The relationship between the canonical ENSO and the phase transition of the Antarctic oscillatio
典型相关分析的spss操作流程
典型相关分析的spss操作流程1.首先，打开SPSS软件并创建一个新的数据文件。
First, open the SPSS software and create a new data file.2.导入你要进行典型相关分析的数据到SPSS中。
Import the data for canonical correlation analysis into SPSS.3.确保数据变量的命名和类型是正确的。
Make sure the data variable names and types are correct.4.确认数据的缺失值情况，并进行适当的处理。
Check for missing values in the data and handle them appropriately.5.选择“分析”菜单中的“相关”选项。
Select the "Correlate" option from the "Analysis" menu.6.选择“典型相关”作为分析的方法。
Choose "Canonical Correlation" as the method for analysis.7.将想要进行分析的自变量和因变量添加到对应的框中。
Add the predictor and criterion variables to their respective boxes for analysis.8.确定是否需要进行变量的标准化处理。
Decide if standardization of variables is needed.9.点击“OK”开始进行典型相关分析。
Click "OK" to start the canonical correlation analysis.10.解释典型相关分析的结果和统计显著性。
Interpret the results and statistical significance of the canonical correlation analysis.11.对典型相关分析的结果进行图表展示。
典型相关分析及其应用实例
摘要典型相关分析是多元统计分析的一个重要研究课题.它是研究两组变量之间相关的一种统计分析方法，能够有效地揭示两组变量之间的相互线性依赖关系.它借助主成分分析降维的思想，用少数几对综合变量来反映两组变量间的线性相关性质.目前它已经在众多领域的相关分析和预测分析中得到广泛应用.本文首先描述了典型相关分析的统计思想，定义了总体典型相关变量及典型相关系数，并简要概述了它们的求解思路，然后深入对样本典型相关分析的几种算法做了比较全面的论述.根据典型相关分析的推理，归纳总结了它的一些重要性质并给出了证明，接着推导了典型相关系数的显著性检验.最后通过理论与实例分析两个层面论证了典型相关分析的应用于实际生活中的可行性与优越性.【关键词】典型相关分析，样本典型相关，性质，实际应用ABSTRACTThe Canonical Correlation Analysis is an important studying topic of the Multivariate Statistical Analysis. It is the statistical analysis method which studies the correlation between two sets of variables. It can work to reveal the mutual line dependence relation availably between two sets of variables. With the help of the thought about the Principal Components, we can use a few comprehensive variables to reflect the linear relationship between two sets of variables. Nowadays It has already been used widely in the correlation analysis and forecasted analysis.This text describes the statistical thought of the Canonical Correlation Analysis firstly, and then defines the total canonical correlation variables and canonical correlation coefficient, and sum up their solution method briefly. After it I go deep into discuss some algorithm of the sample canonical correlation analysis thoroughly. According to the reasoning of the Canonical Correlation Analysis, sum up some of its important properties and give the identification, following it, I infer the significance testing about the canonical correlation coefficient. According to the analysis from the theories and the application, we can achieve the possibility and the superiority from canonical correlation analysis in the real life.【Key words】Canonical Correlation Analysis，Sample canonical correlation，Character，Practical applications目录前言 (1)第1章典型相关分析的数学描述 (2)第2章典型变量与典型相关系数 (3)2.1 总体典型相关 (3)2.2 样本典型相关 (4)2.2.1 第一对典型相关变量的解法 (4)2.2.2 典型相关变量的一般解法 (8)2.2.3 从相关矩阵出发计算典型相关 (9)第3章典型相关变量的性质 (11)第4章典型相关系数的显著性检验 (15)第5章典型相关分析的计算步骤及应用实例 (18)5.1 典型相关分析的计算步骤 (18)5.2 实例分析 (19)结语 (26)致谢 (27)参考文献 (28)附录 (29)前言典型相关分析(Canonical Correlation Analysis ,CCA)作为多元统计学的一个重要部分，是相关分析研究的一个主要内容.典型相关分析不仅其方法本身具有重要的理论意义，而且它还可以作为其他分析方法，如多重回归、判别分析和相应分析的工具，因此在多元分析方法中占有特殊的地位.典型相关的概念是在两个变量相关的基础上发展起来的.我们知道，两个随机变量的相关关系可以用它们的简单相关系数来衡量；一个随机变量与一组随机变量之间的相关关系可以用复相关系数来衡量.但考虑一组随机变量与另一组随机变量的关系时，如果运用两个变量的相关关系，分别考虑第一组每个变量和第二组中每个变量的相关，或者运用复相关关系，考虑一组变量中的每个变量和另一组变量的相关，这样做比较繁琐，抓不住要领.因此，为了用比较少的变量来反映两组变量之间的相关关系，一种考虑的思路就是类似主成分分析，考虑两组变量的线性组合，从这两个线性组合中找出最相关的综合变量，通过少数几个综合变量来反映两组变量的相关性质，这样便引出了典型相关分析.典型相关分析的基本思想是首先在每组变量中找出变量的线性组合，使其具有最大相关性，然后再在每组变量中找出第二对线性组合，使其分别与第一对线性组合不相关，而第二对本身具有最大的相关性，如此继续下去，直到两组变量之间的相关性被提取完毕为止.有了这样线性组合的最大相关，则讨论两组变量之间的相关，就转化为只研究这些线性组合的最大相关，从而减少研究变量的个数.典型相关分析是由Hotelling于1936年提出的.就目前而言，它的理论己经比较完善，计算机的发展解决了典型相关分析在应用中计算方面的困难，成为普遍应用的进行两组变量之间相关性分析技术.如在生态环境方面，用典型相关理论对预报场与因子场进行分析，实现了短期气象预测；借助典型相关，分析了植被与环境的关系；在社会生活领域，应用典型相关分析了物价指标和影响物价因素的相关关系等等.第1章 典型相关分析的数学描述一般地，假设有一组变量p X X X ,,,21 与另一组变量q Y Y Y ,,,21 ，我们要研究这两组变量之间的相关关系，如何给两组变量之间的相关性以数量的描述.当q p 1时，就是我们常见的研究两个变量X 与Y 之间的简单相关关系，其相关系数是最常见的度量，定义为：)()(),(Y Var X Var Y X Cov xy当1 p ,1 q （或1,1 p q ）时，p 维随机向量'21),(p X X X X ，设),(~1p N Y X ， 22211211，其中，11 是第一组变量的协方差阵，12 是第一组与第二组变量的协方差阵，22 是第二组变量的协方差阵.则称221211121R 为Y 与p X X X ,,,21 的全相关系数，全相关系数用于度量一个随机变量Y 与另一组随机变量p X X X ,,,21 的相关系数.当1, q p 时，利用主成分分析的思想，可以把多个变量与多个变量之间的相关化为两个新的综合变量之间的相关.也就是做两组变量的线性组合即X X X X U p p '2211 Y Y Y Y V q q '2211其中，'21),,,(p 和'21),,,(q 为任意非零向量，于是我们把研究两组变量之间的问题化为研究两个变量V U 与之间的相关问题，希望寻求 ，使U ，V 之间最大可能的相关，我们称这种相关为典型相关，基于这种原则的分析方法就是典型相关分析.第2章 典型变量与典型相关系数2.1 总体典型相关设有两组随机变量'21),,,(p X X X X ,'21),,,(q Y Y Y Y ,分别为维维和q p 随机向量，根据典型相关分析的思想，我们用X 和Y 的线性组合X ' 和Y ' 之间的相关性来研究两组随机变量X 和Y 之间的相关性.我们希望找到 和，使得）（‘Y X ', 最大.由相关系数的定义)()(),(),(''''''Y Var X Var Y X Cov Y X易得出对任意常数d c f e ,,,，均有),(])(,)([''''Y X d Y c f X e这说明使得相关系数最大的Y X '', 并不唯一.因此，为避免不必要的结果重复，我们在求综合变量时常常限定1)(' X Var ， 1)(' Y Var于是，我们就有了下面的定义：设有两组随机变量'21),,(p X X X X ，'21),,(q Y Y Y Y ，q p 维随机向量Y X 的均值向量为零，协方差阵0 （不妨设q p ）.如果存在'1111),,(p 和'1111),,(q ，使得在约束条件1)(' X Var ，1)(' Y Var 下，),(m ax ),('''1'1Y X Y X则称Y X '1'1, 是Y X ,的典型相关变量，它们之间的相关系数称为典型相关系数；其他典型相关变量定义如下：定义了前1 k 对典型相关变量之后，第k 对典型相关变量定义为：如果存在'1),,(pk k k 和'1),,(qk k k ，使得 ⑴ Y X k k '', 和前面的1 k 对典型相关变量都不相关；⑵ 1)(' X Var k ，1)(' Y Var k ； ⑶ Y X k k '' 和的相关系数最大，则称Y X k k '' 和是Y X ,的第k 对（组）典型相关变量，它们之间的相关系数称为第k 个典型相关系数（p k ,,2 ）.2.2 样本典型相关以上是根据总体情况已知的情形进行，而实际研究中，总体均值向量 和协方差阵 通常是未知的，因而无法求得总体的典型相关变量和典型相关系数，首先需要根据观测到的样本数据阵对 进行估计. 2.2.1 第一对典型相关变量的解法设总体'11),,,,,(q p Y Y X X Z ，已知总体的n 次观测数据为：1)()()()(q p t t t Y X Z （n t ,,2,1 ）， 于是样本数据阵为)(212122221222211121111211q p n nq n n np n n q p q p y y y x x x y y y x x xy y y x x x若假定),,(~ q p N Z 则由参考文献【2】中定理2.5.1知协方差阵 的最大似然估计为'1)()()()(1nt t t Z Z Z Z n其中Z = nt t Z n 1)(1，样本协方差矩阵S 为：22211211S S S SS 式中nj j j X X X X n S 1'11)()(1'112)()(1 Y Y X X n S j nj j 21S nj j j X X Y Y n 1')()(1 '122)()(1 Y Y Y Y n S j nj jn j j X n X 11， nj j Y n Y 11令j j X U ' ，j j Y V ' ，则样本的相关系数为nj jnj jj nj j j j V VU UV V U U V U r 1212'1)()()()(),(又因为：X X n X n U n U n j j n j j n j j '1'1'1111Y Y n Y n V n V n j j n j j n j j '1'1'111112''''1'''1)()(1)()(1S Y Y X X n V V U U n S j n j j j n j j V U jj 11''''1'''1)()(1)()(1S X X X X n U U U U n S j n j j j n j j U U jj 22''''1'''1)()(1)()(1S Y Y Y Y n V V V V n S j n j j j n j j V V jj 所以22'11'12'),(S S S V U r j j由于j U ，j V 乘以任意常数并不改变他们之间的相关系数，即不妨限定取标准化的j U 与j V ，即限定j U 及j V 的样本方差为1，故有：1 j j j j V V U U S S （2.2.1） 则 12'),(S V U r j j （2.2.2） 于是我们要求的问题就是在（2.2.1）的约束条件下，求p R ，q R ，使得式（2.2.2）达到最大.这是条件极值的问题，由拉格朗日乘子法，此问题等价于求 ， ，使)1(2)1(2),(22'11'12'S S S（2.2.3） 达到最大.式中，，为拉格朗日乘数因子.对上式分别关于 ， 求偏导并令其为0，得方程组：0022211112S S S S （2.2.4）分别用' ，' 左乘方程（2.2.4）得22'21'11'12'S S S S 又 '12')( S 21'S 所以'12'21')(S S也就是说，正好等于线性组合U 与V 之间的相关系数，于是（2.2.4）式可写为：0022211112 S S S S 或 022211211S S S S（2.2.5） 而式（2.2.5）有非零解的充要条件是：022211211S S S S （2.2.6）该方程左端是的q p 次多项式，因此有q p 个根.求解的高次方程（2.2.6），把求得的最大的代回方程组（2.2.5），再求得 和 ，从而得出第一对典型相关变量.具体计算时，因的高次方程（2.2.6）不易解，将其代入方程组（2.2.5）后还需求解q p 阶方程组.为了计算上的方便，我们做如下变换：用12212 S S 左乘方程组（2.2.5）的第二式，则有12212 SS 21S -02212212S S S 即 12212 S S 21S = 12S又由（2.2.5）的第一式，得 1112S S代入上式： 12212 SS 21S 0112S(0)1122112212 S S S S （2.2.7）再用111 S 左乘式（2.2.7），得(111S12212 SS 0)221p I S （2.2.8）因此，对2有p 个解，设为22221p r r r ，对 也有p 个解.类似地，用11121 S S 左乘式（2.2.5）中的第一式，则有011111211211121S S S S S S （2.2.9）又由（2.2.5）中的第二式，得2221S S代入到（2.2.8）式，有 11121( SS 12S 0)222S再以122 S 左乘上式，得0)(21211121122q I S S S S （2.2.10）因此对2有q 个解，对 也有q 个解，因此2为111S 12212 S S 21S 的特征根， 是对应于2的特征向量.同时2也是1211121122S S S S 的特征根， 为相应特征向量.而式（2.2.8）和（2.2.10）有非零解的充分必要条件为：002121112112222112212111q p I S S S S I S S S S （2.2.11）对于（2.2.11）式的第一式，由于011 S ，022 S ，所以0111S ，0122 S ，故有：2112212111S S S S 2121221221221112111S S S S S S 而2121221221221112111S S S S S S 与2111211222122122111 S S S S S S 有相同的特征根.如果记T 12212111 S S S则 2111211222122122111S S S SS S='T T类似的对式（2.2.11）的第二式，可得T T S S SSS S'21221221112111212122而'T T 与T T '有相同的非零特征根，从而推出（2.2.8）和（2.2.10）的非零特征根是相同的.设已求得'T T 的p 个特征根依次为： 022221p则T T '的q 个特征根中，除了上面的p 个外，其余的p q 个都为零.故p 个特征根排列是021 p ，, 1210 p p ，因此，只要取最大的1 ，代入方程组（2.2.5）即可求得相应的1 ，1 .令U =X '1 与Y V '1 为第一对典型相关变量，而1'112'1),( S V U r 为第一典型相关系数.可见求典型相关系数及典型相关变量的问题，就等价于求解'T T 的最大特征值及相应的特征向量. 2.2.2 典型相关变量的一般解法从样本典型相关变量的解法中，我们知道求典型相关变量和典型相关系数的问题，就是求解'T T 的最大特征值及相应的特征向量.不仅如此，求解第k 对典型相关变量和典型相关系数，类似的也是求'T T 的第k 大的特征值和相应的特征向量.下面引用参考文献【2】中定理10.1.1 来得出样本典型相关的一般求法.设总体的n 次观测数据为：1)()()()( q p t t t Y X Z （n t ,,2,1 ） 不妨设q p ，样本均值为0，协方差矩阵S 为：22211211S S S SS 0 记2122122111S S ST ，并设p 阶方阵'T T 的特征值依次为022221p （p i i ,,1,0 ）；而p l l l ,,,21 为相应的单位正交特征向量.令 kk l S2111，k k k S S 211221则X U k k '，Y V kk '为Y X ,第k 对典型相关变量，'k为第k 典型相关系数. 由上述分析不难看出，典型相关系数i 越大说明相应的典型变量之间的关系越密切，因此一般在实际中忽略典型相关系数很小的那些典型变量，按i 的大小只取前n 个典型变量及典型相关系数进行分析. 2.2.3 从相关矩阵出发计算典型相关以上我们从样本协方差阵S 出发，导出了样本典型相关变量和样本典型相关系数.下面我们从样本相关阵R 出发来求解样本典型相关变量和样本典型相关系数.设样本相关阵为)(ij r R ，其中jj ii ij ij s s s r / ，ij s 为样本协方差阵S 的i 行j 列元素.把R 相应剖分为22211211R R R R R 有时，Y X 和的各分量的单位不全相同，我们希望在对各分量作标准化变换之后再做典型相关.记)(1X E ，)(2Y Epp s s D 00111q p q p p p s s D ,1,1200则 111111D R D S ，222222D R D S 212112D R D S ，121221D R D S , 对Y X 和的各分量作标准化变换，即令)(111* X D X ，)(212* Y D Y现在来求*X 和*Y 的典型相关变量*'*X i ，*'*Y i ，m i ,,2,1 . **11111111X X S D S D R**11222222Y Y S D S D R **11112212X Y S D S D R **11221121Y X S D S D R于是1121122121111112112112221212121111111112112212111)()( D S S S S D D S D D S D D S D D S D R R R R因为 2112212111S S S S i i i r 2 1121122121111 D S S S S D )()(121i i i D r D 所以 2112212111R R R R *2*i i i r 式中*i i D 1 ，有111'1111'*11'* i i i i i i S D R D R同理： 1211121122R R R R *2*i i i r 式中*i i D 1 ，有122'2222'*22'* i i i i i i S D R D R ，由此可见*i ，*i 为**,Y X 的第i 对典型系数，其第i 个典型相关系数为i r ，在标准化变换下具有不变性.第3章 典型相关变量的性质根据典型相关分析的统计思想及推导，我们归纳总结了典型相关变量的一些重要性质并对总体与样本分别给出证明.性质1 同一组的典型变量互不相关 ⅰ总体典型相关设Y X 与的第i 对典型变量为X U i i ' ，Y V i i ' ，m i ,,2,1则有 0),( j i U U 0),( j i V V m j i 1 证明详见参考文献【5】. ⅱ样本典型相关设Y X 与的第i 对典型变量为X U i i ' ，Y V i i ' ，m i ,,2,1因为 '111i i U U i i S S ，'221i iVV i i S S ，m i ,,2,1 '11(,)0i j i j U U i j r U U S S ，m j i 1'22(,)0i ji j VV i j r V V S S ，m j i 1 表明由X 组成的第一组典型变量m U U U ,,,21 互不相关，且均有相同的方差1；同样，由Y 组成的第二组典型变量m V V V ,,,21 也互不相关，且也有相同的方差1.性质2 不同组的典型变量之间的相关性ⅰ总体典型相关i i i V U ),( m i ,,2,10),( j i V U m j i 1 证明详见参考文献【5】. ⅱ样本典型相关i i i i i r V U r S ),(12' ， m i ,,2,1'1211''22111222(,)0,1i j i j U V i ji j j i j r U V S S S S S r i j m表明不同组的任意两个典型变量，当j i 时，相关系数为i r ；当j i 时是彼此不相关的.记'21),,,(m U U U U ，'21),,,(m V V V V ，则上述性质可用矩阵表示为 ,UU m VV m S I S IUV S或 mm IU S I V其中12(,,...,)m diag r r r性质3 原始变量与典型变量之间的关系 求出典型变量后，进一步计算原始变量与典型变量之间的相关系数矩阵，也称为典型结构.下面我们分别对总体与样本进行讨论.ⅰ总体典型相关的原始变量与典型变量的相关性详见参考文献【2】. ⅱ样本典型相关 记m p ij m A )(),,,(21 m q ij m B )(),,,(21S22211211S S S S =q p q p p q p pq p q p q p p p p p p p q p p p p pp p q p p p s s s s s s s s s s s s s s s s ,1,,1,,11,1,11,1,1,1,11,1111则A S X A X A X X n S n i i XU11'''1)()(1 B S X B X B X X n S n i i XV12'''1)()(1 A S X A X A Y Y n S n i i YU21'''1)()(1 B S Y B Y B Y Y n S n i i YV22'''1)()(1所以利用协方差进一步可以计算原始变量与典型变量之间的相关关系.若假定原始变量均为标准化变量，则通过以上计算所得到的原始变量与典型变量的协方差阵就是相关系数矩阵.1(,)pi j ik k r X U s,1(,)qi j i p k k r X V sp i ,,2,1 ， m j ,,2,1,1(,)pi j i p k kjk r Y U s,1(,)qi j i p p k kjk r Y V s q i ,,2,1 ， m j ,,2,1性质4 设Y X 和分别为维维和q p 随机向量，令d X C X '*，h Y G Y '*，其中C 为p p 阶非退化矩阵，d 为p 维常数向量，G 为q q 阶非退化矩阵，q h 为维常数向量.则：ⅰ对于总体典型相关有：⑴ **Y X 和的典型相关变量为*'*)(X a i 和*'*)(Y b i ，其中i i a C a 1* ，i i b G b 1* （p i ,,2,1 ）；而i i b a 和是Y X 和的第i 对典型相关变量的系数.⑵ ],[])(,)[(''*'**'*Y b X a Y b X a i i i i ，即线性变换不改变相关性. 证明详见参考文献【2】.ⅱ对于样本典型相关有：⑴ **Y X 和的典型相关变量为*'*)(X a i 和*'*)(Y b i ，其中i i a C a 1* ，i i b G b 1* （p i ,,2,1 ）；而i i b a 和是Y X 和的第i 对典型相关变量的系数.⑵ ],[])(,)[(''*'**'*Y b X a r Y b X a r i i i i ，即线性变换不改变相关性. 证明：⑴ 设**Y X 和的典型相关变量分别为*'*)(X a U i ，*'*)(Y b V i由于 i i a C a 1* ，i i b G b 1*d X C X '*，h Y G Y '*所以 d C a X a d X C C a d X C a C U i i i i '1''''1'''1)()()()()(h G b Y b h Y G G b h Y G b G V i i i i '1''''1'''1)()()()()(即有i i b a 和是Y X 和的第i 对典型相关变量的系数. ⑵ 由⑴的证明可知*'*)(X a U i d C a X a i i '1'')( *'1'''*)()(h G b Y b Y b V i i i由于d C a i '1')( 与h G b i '1')( 都是常数，所以],[])(,)([])(,)[('''1'''1''*'**'*Y b X a r h G b Y b d C a X a r Y b X a r i i i i i i i i 即有线性变换不改变相关性.性质5 简单相关、复相关和典型相关之间的关系当1 q p ， Y X 与之间的（惟一）典型相关就是它们之间的简单相关；当Y X q p 与时或,11 之间的（惟一）典型相关就是它们的复相关.复相关是典型相关的一个特例，而简单相关又是复相关的一个特例.从第一个典型相关的定义可以看出，第一个典型相关系数至少同)(Y X 或的任一分量与)(X Y 或的复相关系数一样大，即使所有这些复相关系数都很小，第一个典型相关系数仍可能很大；同样，从复相关的定义也可以看出，当1 p （或1 q ）时，)()(X Y Y X 或与或之间的复相关系数也不会小于)()(X Y Y X 或与或的任一分量之间的相关系数，即使所有这些相关系数都很小，复相关系数仍可能很大.第4章 典型相关系数的显著性检验设总体Z 的两组变量'21),,,(p X X X X ，'21),,,(q Y Y Y Y ，且'),(Y X Z ),(~ q p N ，在做两组变量X ，Y 的典型相关分析之前，首先应该检验两组变量是否相关，如果不相关，则讨论两组变量的典型相关就毫无意义. 1．考虑假设检验问题：0H ：021 m1H ：m ,,,21 至少有一个不为零其中 q p m ,m in .若检验接受0H ，则认为讨论两组变量之间的相关性没有意义；若检验拒绝0H ，则认为第一对典型变量是显著的.上式实际上等价于假设检验问题0H ：0),(12 Y X Cov ， 1H ：012用似然比方法可导出检验0H 的似然比统计量||||||2211S S S其中q p 阶样本离差阵S 是 的最大似然估计，且S =22211211S S S S ，11S ，22S 分别是11 ，22 的最大似然估计.该似然比统计量 的精确分布已由霍特林（1936），Girshik （1939）和Anderson （1958）给出，但表达方式很复杂，又不易找到该分布的临界值表，下面我们采用 的近似分布.利用矩阵行列式及其分块行列式的关系，可得出：||·||||21122121122S S S S S S =|S S S S |·|S |·||21-12212-1111122 p S所以)1(001001||212212112212111ipi p p S S S S其中 2i是'TT 的特征值（2122122111S S S T ），按大小次序排列为 2122 02 p，当1 n 时，在0H 成立下 ln 0m Q 近似服从2f 分布，这里pq f ，)1(211 q p n m ，因此在给定检验水平 之下，若由样本算出的20 Q 临界值，则否定0H ，也就是说第一对典型变量1 U ，1V 具有相关性，其相关系数为1 ，即至少可以认为第一个典型相关系数1为显著的.将它除去之后，再检验其余1 p 个典型相关系数的显著性，这时用Bartlett 提出的大样本2 检验计算统计量：pi ip22223221)1()1()1)(1(则统计量11ln )]1(212[ q p n Q近似地服从（1 p ）（1 q ）个自由度的2分布，如果21 Q ，则认为2显著，即第二对典型变量2U ，2V 相关，以下逐个进行检验，直到某一个相关系数k检验为不显著时截止.这时我们就找出了反映两组变量相互关系的1 k 对典型变量.2．检验)(0k H ： ),,2(0p k k当否定0H 时，表明Y X ,相关，进而可以得出至少第一个典型相关系数01 ，相应的第一对典型相关变量11,V U 可能已经提取了两组变量相关关系的绝大部分信息.两组变量余下的部分可认为不相关，这时0 k ),,2(p k ，故在否定0H 后，有必要再检验)(0k H ),,2(p k ，即第k 个及以后的所有典型相关系数均为0),,3,2(p k .为了减少计算量，下面我们采用二分法来减少检验次数，取检验统计量为p ki i k q p k n Q )1ln()]1(21[2它近似服从)1)(1( k q k p 个自由度的2 分布.在检验水平 下，若)]1)(1[(2k q k p Q k ，则拒绝0H ，即认为第k 对典型相关系数在显著性水平 下是显著的，否则不显著.从第2个典型相关系数到第p 个典型相关系数，共1 p 个数，所以根据二分法的原理，将它们分为一个区间 p ,2，然后先检验第 21p 个典型相关系数即中位数，当021p 时，即认为第 21p 个典型相关系数不相关，否定原假设，接着检验21,2p ；若当021p 时，则检验p p ,21.如此划分区间依次检验下去，由数学分析上的区间套定理，一定存在第k 个数),,3,2(p k ，使得01 k ，而0 k .以上的一系列检验实际上是一个序贯检验，检验直到对某个k 值0H 未被拒绝为止.事实上，检验的总显著性水平已不是 了，且难以确定.还有，检验的结果易受样本容量大小的影响.因此，检验的结果只宜作为确定典型变量个数的重要参考依据，而不宜作为惟一的依据.第5章 典型相关分析的计算步骤及应用实例5.1 典型相关分析的计算步骤设)()1(,,n X X 为取自正态总体的样本（实际上，相当广泛的情况下也对），每个样品测量两组指标，分别记为'1),,(p X X X ，'1),,(q Y Y Y ，原始资料矩阵为：)(212122221222211121111211q p n nq n n np n n q p q p y y y x x x y y y x x xy y y x x x第一步 计算相关矩阵R ，并将R 剖分为22211211R R R R R 其中11R ，22R 分别为第一组变量和第二组变量之间的相关系数矩阵，'2112R R 为第一组与第二组变量之间的相关系数.第二步 求典型相关系数及典型变量首先求2112212111R R R R A的特征根 2i，特征向量)(1i D；1211121122R R R R B的特征根2i，特征向量)(2i D.)()(111)(i i D D，)()(212)(i i D D写出样本的典型变量为 X U ’)1(1，Y V ’)1(1X U ’)2(2，Y V ’)2(2X U p p ’)(，Y V p p ’)(第三步 典型相关系数的显著性检验 首先，检验第一对典型变量的相关系数，即0H ：0^1 ，1H ：0^1它的似然比统计量为pi i p1^2^2^22^211)1()1()1)(1(则统计量11ln )]1(212[ q p n Q给定显著性水平 ，查表得2，若21 Q ，则否定0H ，认为第一对典型变量相关，否则不相关.如果相关则依次逐个检验其余典型相关系数，直到某一个相关系数^k ),,2(p k 检验为不显著时截止.5.2 实例分析例1：某康复俱乐部对20名中年人测量了三个生理指标：体重)(1x 、腰围（2x ）、脉搏（3x ）和三个训练指标：引体向上（1y ）、起坐次数（2y ）、跳跃次数（3y ）.数据如附录1：解：记'321),,(x x x X ，'321),,(y y y Y ，其中样本容量20 n .附录1中的数据用SPSS 统计软件计算得六个变量之间的相关矩阵如下：n Sig.(2-tailed) .113 .127. .526 .340 .884 N 20 20 20 202020 Y1Pearson Correlatio n -.390 -.552(*) .1511 .696(**).496(*)Sig.(2-tailed) .089 .012.526 . .001 .026 N 20 20 20202020Y2PearsonCorrelatio n -.493(*)-.646(**).225 .696(**) 1 .669(**)Sig.(2-tailed) .027 .002.340 .001 . .001 N 20 20 20 202020 Y3Pearson Correlatio n -.226 -.191 .035.496(*) .669(**)1Sig.(2-tailed) .337 .419.884 .026 .001 . N 20 2020202020** Correlation is significant at the 0.01 level (2-tailed).* Correlation is significant at the 0.05 level (2-tailed).即样本相关矩阵为：11R =1353.0366.01870.0122R =1669.0496.01696.01'2112R R =035.0225.0151.0192.0646.0552.0226.0493.0390.0于是特征方程 022112212111 R R R R用Matlab 求得矩阵2112212111R R R R 的特征值分别为0.6630、0.0402和0.0053，于是 797.01 ，201.02 ，073.03下面我们进行典型相关系数的显著性检验，先检验第一对典型变量的相关系数，欲检验：0H ：01 ， 1H ：01 它的似然比统计量为)1)(1)(1(2322211 =3504.0)0053.01)(0402.01)(6330.01( 255.163504.0ln 5.15ln )]333(2120[11 Q查2 分布表得，919.16)9(205.0 ，因此在05.0 的显著性水平下，)9(205.01 Q ，所以拒绝原假设0H ，也即认为第一对典型相关变量是显著相关的.然后检验第二对典型变量的相关系数，即进一步检验：0H ：02 ， 1H ：02它的似然比统计量为9547.0)0053.01)(0402.01()1)(1(23222 )4(488.9745.09547.0ln 08.16ln ])333(21120[205.02212 Q 所以无法否定原假设0H ，故接受0H ：02 ，即认为第二对典型相关变量不是显著相关的.由以上检验可知只需求第一对典型变量即可. 于是求797.01 的特征向量 *1，而*1*12112211R R ，解得059.0579.1775.0*1，716.0054.1350.0*1 ， 因此，第一对样本典型变量为*3*2*1*1059.0579.1775.0x x x u *3*2*1*1716.0054.1350.0y y y vY X 与第一对典型变量的相关系数为797.01 ，可见两者的相关性较为密切，即可认为生理指标与训练指标之间存在显著相关性.例2：为了研究某企业不同部门人员工作时间的关系，随机选取25个企业进行入户调查，达到25个被访企业业务部门和技术部门经理每月工作时间和员工每月工作时间（单位为小时），具体数据如附表2分析：设业务部门经理和员工每月工作时间为（21,X X ），技术部门经理和员工每月工作时间为（21,Y Y ），利用典型相关分析研究企业业务部门和技术部门人员工作时间的关系.解：样本容量为25 n ，2 p ，2 q 分别为随机变量Y X 与的维数.⑴ 标准化随机变量'21),(X X X 与'21),(Y Y Y .根据样本均值i x与标准差ii S ，依照公式iiiki ki S x x x*，对数据标准化.⑵ 求解Y X 的相关矩阵R ，并将其分块yy yxxy xx R RR R R . 将数据输入SPSS 软件求得相关系数矩阵如下：Correlations** Correlation is significant at the 0.01 level (2-tailed).所以样本相关矩阵1834.0705.0705.01693.0711.01735.01R 分块后2222 yy yx xy xx R RR R R ⑶ 求解534949.0538840.0538840.0544309.011111yx yy xy xx R R R R M 的两个非零特征根，解得两个非零特征根为6218.021 ，0029.022 .⑷ 进行相关系数的显著性检验，取r m 个显著性检验不为0的特征根.Y X 与第一对典型变量的相关系数为7885.01 ，Y X 与第二对典型变量的相关系数为0537.02 .先检验第一对典型变量的相关系数，假设01H ：01 （即第一对典型变量不相关），由典型相关系数的值可得3771.0)1)(1(22211计算统计量97.203771.0ln )5.224(ln )]1(21)1[(11 q p n Q 对于给定的显著性水平05.0488.9)4()1)(1(97.20205.021 m q m p Q所以否定零假设.01H ：01 ，即第一对典型变量是显著相关的.然后检验第二对典型变量的相关系数，假设02H ：02 （即第二对典型变量不相关），由典型相关系数的值可得9971.0)1(222 计算统计量05945.09971.0ln )5.224(ln )]1(21)2[(22 q p n Q 对于给定的显著性水平05.0841.3)1()1)(1(05945.0205.022 m q m p Q所以无法否定假设.02H ：02 ，即第二对典型变量不是显著相关的.由以上检验可知，只需求第一对典型变量即可.⑸ 求1 m 个显著性检验不为0的特征根21 的特征向量1l ，而11111l R R m yx yy，解得'1)521548.0,55216.0( l ，'1)538134.0,504018.0( m .⑹ 求出r 对典型相关变量X l u j j ' ，Y m v j j ' ，.,,2,1m j 根据上面求得的特征向量11m l 和，得第一对典型相关变量为21'1121'11538134.0504018.0521548.055216.0Y Y Y m v X X X l u Y X 与第一对典型变量的相关系数为7885.01 ，可见其相关性较为密切.⑺ 由于21'11521548.055216.0X X X l u ，与业务部门经理和员工每月工作时间都成正比，而且系数差不多，所以u可以解释为业务部门人员工作时间.同1理v可以解释为技术部门人员的工作时间.可见一个企业技术部门和业务部门人1员月工作时间存在显著的相关性.典型相关分析是一种采用类似主成分分析的做法，在每一组变量中都选择若干个有代表性的综合指标（变量的线性组合），通过研究两组的综合指标之间的关系来反映两组变量之间的相关关系.在实际中，只须着重研究相关关系较大的那几对典型相关变量.本文首先根据典型相关分析的统计理论，初步探讨了总体典型相关变量和典型相关系数，然后重点讨论了样本典型相关分析，以及它们的一系列性质与显著性检验，并做了相应的实例分析.通过实例分析，我们进一步明确了典型相关分析是研究两组变量之间相关性的一种降维技术的统计分析方法.而复相关是典型相关的一个特例，简单相关是复相关的一个特例.第一对典型相关包含有最多的有关两组变量间相关的信息，第二对其次，其他对依次递减.各对典型相关变量所含的信息互不重复.并且经标准化的两组变量之间的典型相关系数与原始的两组变量间的相应典型相关系数是相同的.本文是在我的指导老师吴可法教授的精心指导和悉心关怀下完成的，在我的学习生涯和论文工作中无不倾注着老师的辛勤汗水和殷切关怀.吴老师宽厚的人格、敏捷的思维、严谨的治学态度、渊博的知识、积极向上的人生态度、平易近人的师长风范和两年来的谆谆教导，使我深受启迪，并永远铭记在心.从吴老师身上，我不仅学到了扎实的专业知识和技能，更学到了做人的道理，这些教诲必将成为惠及一生的宝贵财富.在此谨向吴老师致以最衷心的感谢和美好的祝愿!论文期间，我得到了许多老师和同学的帮助，本人在这里对他们致以衷心的感谢.我还要感谢我的家人，是他们的理解、支持和鼓励，使我的学习能够顺利进行.最后衷心感谢在百忙之中评审论文和参加答辩的各位专家、教授!。
典型相关分析方法研究
典型相关分析方法研究摘要：典型相关分析是研究两组变量（或两个随机向量）之间的相关关系的一种统计方法。
与仅研究二个变量间线性关系的简单相关分析相比,典型相关分析能揭示出两组变量之间的内在联系,且两组变量的数目可以改变,这确定了它的重要性。
随着计算机技术的发展,典型相关分析在各个行业试验研究中应用日渐广泛.本文主要介绍典型相关分析的基本原理与步骤并举例说明其应用.关键词:典型相关分析;基本原理；步骤；应用Abstract:Canonical correlation analysis is the study of two groups of variables （or two random vectors）a statistical method the relationship between the. Compared with only the simple correlation analysis of linear relationship between two variables and canonical correlation analysis can reveal the internal relations between two sets of variables，and the number of two groups of variables can change，this determines the importance of it. With the development of computer technology, the canonical correlation analysis system has been widely used in various industries in experimental study。
This paper mainly introduces the basic principle and procedure of canonical correlation analysis and examples of its application.Key words：Canonical correlation analysis; basic principle；step; application一、引言典型相关分析（Canonical Correlation Analysis 简称CCA)是处理两个随机矢量之间相关性的统计方法,在多元统计分析中占有非常重要的地位。
生态批评
生态批评是一个言人人殊的话语体。
大多数人认同彻丽尔·格罗特费尔蒂的定义：“生态批评是探讨文学与自然环境之关系的批评。
”一般认为，“生态批评”这一概念由美国学者威廉·鲁克尔曼1978年首次提出，他的《文学与生态学：一次生态批评实验》文章在《衣阿华评论》1978冬季号上刊出，以“生态批评”概念明确地将“文学与生态学结合起来”。
1992年，“文学与环境研究会”在美国内华达大学成立。
1994年，克洛伯尔出版专著《生态批评：浪漫的想象与生态意识》，提倡“生态学的文学批评”(ecological literary criticism)或“生态学取向的批评”(ecological oriented criticism)。
1995年在科罗拉多大学召开了首次研讨会，会议部分论文以《阅读大地：文学与环境研究的新走向》为书名正式出版（1998）。
其后，生态批评的著作有如雨后春笋般地充斥文论界。
[3]1996年美国第一本生态批评论文集《生态批评读本》由格罗特费尔蒂和弗罗姆主编出版，其宗旨在于“分别讨论生态学及生态文学理论、文学的生态批评和生态文学的批评”，使得生态批评更具有文学批评的特征和范式。
在导言中格罗特费尔蒂（Cheryll Glotfelty）给生态批评加以定义：“生态批评研究文学与物理环境之间的关系。
正如女性主义批评从性别意识的视角考察语言和文学，马克思主义批评把生产方式和经济阶级的自觉带进文本阅读，生态批评运用一种以地球为中心的方法研究文学。
”1998年英国第一本生态批评论文集《书写环境：生态批评与文学》在伦敦出版，分生态批评理论、生态批评的历史、当代生态文学三个部分。
这本由克里治和塞梅尔斯主编的著作认为：“生态批评要探讨文学里的环境观念和环境表现”。
1999年夏季的《新文学史》是生态批评专号，共发表十篇专论生态批评的文章，2000年出版的生态批评著作主要有默菲教授主编的论文集《自然取向的文学研究之广阔领域》，托尔梅奇等主编的《生态批评新论集》，贝特的《大地之歌》等。
Canoncial+Correlation+Analysis
Canoncial Correlation Analysis on the Relationship Between China's Transportation Industry and National Economic DevelopmentLI WeidongSchool of Economics & Management,Beijing Jiaotong University, P.R.China,100044 Abstract: Canonical correlation analysis method is used to analyze the quantitative relationship between China’s transportation industry and national economic development. At first, literature overview for the research on the relationship between transportation industry and national economic development is taken. Then as the result of the complexity of transportation industry and national economy system, the index systems that reflect the development of transportation and economic development are designed. The transportation index system includes the indicators of transport turnover volume. Economic development indicators consist of gross national product of different industries. Then, based on data collection of above indicators from 1952 to 2007, canoncial correlation analysis method is used to analyze the relationship between the transportation and national economic development. The results show that there are strong correlation between the China’s transportation industry and national economic development. The highway traffic and railway traffic have contributed greatly to the national economic development.Key words: national economic development, canoncial correlation analysis,transportation industry1 IntroductionTransportation industry is one of the important basic industries in national economy. Undoubtedly highway transportation industry plays an important role in national economic development. Domestic and foreign scholars have studied on the relationship between highway transportation industry and national economic development and obtained some useful conclusions.At first, the research of relationship between the transportation industry and national economic development mainly focused on the highway transport's contribution to national economic development. From the traditional view, transportation has a strong impact on the development of the national economy. With the improvement of transport conditions national production growth is promoted greatly. The positive relationship between transportation and national economic development can be divided into direct transport input effect and the indirect effect which includes multiplier effect. Good transportation conditions provide a lower cost of transportation. So the market scope and scale of production are expanded[1]. The indirect effects comes from the transport infrastructure construction and the creation of employment opportunities in transportation services operation. In addition, it includes multiplier effect of huge investment in construction of a modern transport system in need of steel, wood, cement,etc. Therefore, transportation is a necessary condition but not sufficient condition to economic development.In China many research about the relationship between the transportation and national economy has been conducted, such as Huang Zhendong(1993)[2],Rong Chaohe(1993)’s transportatioinalized theory[3], Hanbiao(1994) ‘s push-pull theory[4] and Xiong Rongjun(1997)’s transportation cost threshold theory[5]. Wang Chuanxu(2004) discussed on the contribution of transportation to national economy[6]. Wang Libin，Wu Qunqi(2006) discussed on the contribution of highway investment to national economy[7]. Domestic and foreign scholars mainly focus on the qualitative relationship between transportation industry and national economic development. But the quantitative research is still a little. In this paper canoncial correlation method is used to analyze the relationship between the transportation industry and national economic development.2 Principles of Canoncial correlation AnalysisCanonical correlation analysis is developed by H.Hotelling in 1936. Canonical correlation analysis (CCA) is an extension of examining the maximum correlation relationship between the dependent variable and a linear combination of the independent variables. Canonical correlations measure the association between two sets of two or more independent variables[8].Canonical correlations are actually the maximum correlation between a linear combination of the dependent variables and a linear combination of the independent variables.It conducts a multiple regression with each set of variables and so creates a linear combination or variate – which is called a canonical variate - for each of the two sets.The technique uses an iterative or repeating process to create the canonical variates for each set of variables, with the goal of finding the correlation of the two canonical variates that is as large as possible. That is, it calculates a ‘starting guess’ version of the linear variate for each set of variables, correlates the two variates, then adjusts the canonical loadings in one or both variates and calculates the correlation again, then adjusts, then calculates, etc., until it finds the loadings that generate the linear variates for each set that result in the highest possible correlation, i.e., when no more adjustments will raise that number.3 Empirical analysis on relationship between China's transportation industry and national economic development3.1Data sourcesThere are five transportation modes which include railway, highway,waterway, civil avaition and pipeline. Because the data of pipeline data is lost in the period of 1952-1970, so we focused on the railway, highway,waterway, civil avaition transportation. To reflect the relationship between transportation industry and national economic development, the passenger traffic turnover volume and freight traffic turnover volume of each transportation mode are chosen to describe the basic development of transportation situation. These index system includes the railway passenger traffic turnover volume(X1), highway passenger traffic turnover volume (X2),waterway passenger traffic turnover volume(X3), civil avaition passenger traffic turnover volume(X4) , railway freight traffic turnover volume (X5), highway freight traffic turnover volume (X6),waterway freight traffic turnover volume (X7), civil avaition freight traffic turnover volume (X8) .There are many indice to reflect the national economic development. GDP of different industry is chosen to reflect the development of national economy. The index system includes GDP of primary industry (Y1), GDP of manufacturing industry (Y2), GDP of construction industry (Y3), GDP of tertiary industry (Y4).The time period is from 1952 to 2007. Here GDP is adjusted with the consumer price index of 1978 as the base period (GDP). Data sources is from the "Summary of New China 50 years’ Statistical Data" and 2005-2008 "China Statistical Yearbook." Freight traffic turnover volume of (unit: 100 million t · km) refers to the total sum of cargo weight multiply by its corresponding distance in a certain period. Passenger traffic turnover volume(unit: 100 million person kilometers) refers to the total sum of passenger number multiply by its corresponding distance in a certain period. The turnover considers the traffic flow and transport distance. It can provide comprehensive understanding of China's transportation industry.3.2Explorative analysis with canoncial correlation analysisAt first, canoncial correlation analysis is taken to the total period (1952-2007).Considering the great difference between the development periods of 1952-1977 and after 1978, then the analysis is taken in the two different periods seperately. Based on the SAS procedure CANCORR, canoncial correlation analysis is taken for the two sets of indice to reflect the transportation and national economy.3.2.1Total period (1952-2007)The results for the canonical correaltions are shown in Table 1. It can be found that the first three pair canoncial correlation coefficient are rather high. The p-value for the canoncial correlationcoefficient are all less than 0.05. The test results indicate that the correlation coeffcient of the four pairof canoncial variate are all significant. Obviously the values of the first three canoncial correlationcoefficients is near to 1. Through the Table 1 it is also found that first pair of canoncial variate canexplain the over 94.8% proportion of variance of data sets. So the first pair canoncial correlation variatesis representative. That shows the canoncial correlation analysis results is rather satisfactory.Table 1 Canonical correlations and function testCorrelations Eigenvalue Difference Proportion Cumulative Pr 10.9976 211.7601 204.7338 0.9483 0.9483 <.0001 20.9356 7.0263 3.1965 0.0315 0.9798 <.0001 30.8904 3.8297 3.1449 0.0172 0.9969 <.0001 4 0.6375 0.6848 0.0031 1.0000 0.5935 0.0001Null hypothesis: remaining correlations are zero.The first pair of canoncial functions is constructed as follows:112345670.410 2.1210.0590.0320.2960.540.3830.135U X X X X X X X =−+−++−−−112341.8440.9250.118 2.034V Y Y Y Y =−+−+8X (1)From the standardized coefficients of the first pair of canoncial functions, it is found that themain factors of transportation is highway passenger traffic turnover volume (X 2), railway freight trafficturnover volume (X 5). The main factor of national economy is GDP of tertiary industry (Y 4), GDP ofmanufacturing industry (Y 2). The results show that development of tertiary industry is mainly influencedby the railway freight traffic, highway passenger traffic.Table 2 Correlations Between the original y variables and the canonical variablesy 1 y 2 y 3 y 4 V 10.9873 0.984 0.9682 0.9958 U 1 0.985 0.9817 0.966 0.9934Table 3 Correlations Between the original x variables and the canonical variablesx 1 x 2x 3 x 4 x 5 x 6 x 7 x 8 V 1 0.94420.98830.10450.94120.93410.97040.9352 0.9434 U 1 0.9420.986 0.10430.939 0.93190.96810.933 0.9412From the results of Table 2 and 3, canonical structure analysis can be taken as follows.From thecanonical loadings of the canoncial function, it is found that the U 1 is highly correlated with all Xvariables except X 3. V 1 is highly positively correlated with all Y variables. Also it is found that V 1 ishighly correlated with all X variables except X 3. U 1 is highly positively positively correlated with all Yvariables. So the first pair of canoncial function can represent the original data sets most effectively.Because the first pair of canoncial variate can explain the variance of the data most efficently, themain factors of transportation industry includes highway passenger traffic turnover volume (X 2), railwayfreight traffic turnover volume (X 5).The main factors of national economy includes GDP of tertiaryindustry (Y 4) and GDP of manufacturing industry (Y 2). Above results show that there are strongcorrelation between the China’s transportation industry and national economic development.3.2.2Planning economy period (1952-1977)The results for the canonical correaltions are shown in Table 4. It can be found that the first twopair canoncial correlation coefficient are rather high. The p-value for the canoncial correlation coefficient are all less than 0.05. The test results indicate that the correlation coeffcient of the two pairsof canoncial variate are all significant. Obviously the values of the first two canoncial correlationcoefficients is near to 1. Through the Table 4 it is also found that first pair of canoncial variate canexplain 90.9% proportion of variance of data sets. So the first pair of canoncial correlation variable ischosen to represent the relationship between transportation and national economy. That shows thecanoncial correlation analysis results is rather satisfactory.Table 4 Canonical correlations and function testCorrelations Eigenvalue Difference Proportion Cumulative Pr1 0.989145.0953 41.7252 0.9099 0.9099 <.0001 2 0.87823.3701 2.6277 0.068 0.9779 0.0038 3 0.65270.7423 0.3883 0.015 0.9929 0.2037 4 0.5113 0.354 0.0071 1 0.7385 0.3491Null hypothesis: remaining correlations are zero.The first pair of canoncial functions is constructed as follows:112345670.479 1.5160.4170.26 1.022 1.900.0250.524U X X X X X X X =+−−+−−+112340.265 1.1010.0720.042V Y Y Y Y =−+−+8X (2)From the standardized coefficients of the first pair of canoncial functions, it is found that the mainfactors of transportation is highway passenger traffic turnover volume (X 2), railway freight trafficturnover volume (X 5). The main factor of national economy is GDP of manufacturing industry (Y 2). Theresults show that development of manufacturing industry is mainly influenced by the highway passengertraffic , railway freight traffic.Table 5 Correlations Between the original y variables and the canonical variablesy 1 y 2y 3 y 4 V 1 0.5985 0.979-0.8086 0.5166 U 1 0.5920.9684 -0.7998 0.5109Table 6 Correlations Between the original x variables and the canonical variablesx 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 U 10.8732 0.9295 0.88910.6730.91220.86040.8203 0.8404 V 1 0.8637 0.9193 0.87940.66560.90230.8510.8113 0.8313Based on the results of Table 5 and 6, it is found that the U 1 is highly positively correlated with allX variables. V 1 is positively correlated with Y 1,Y 2 and Y 4 and negatively correlated with Y 3. Also it isfound that V 1 is positively correlated with all X variables. U 1 is highly positively correlated withY 2 ,Y 1,Y 4 and negatively correlated with Y 3. So the first pair of canoncial function can represent theoriginal data sets most effectively.3.2.3Reform and opening-up period (1978-2007)The results for the canonical correaltions are shown in Table 7. It can be found that the first threepair canoncial correlation coefficient are rather high. The p-value for the canoncial correlation coefficient are all less than 0.05. The test results indicate that the correlation coeffcient of the first threepairs of canoncial variate are all significant. Obviously the values of the first three canoncialcorrelation coefficients is near to 1. Through the Table 1 it is also found that first pair of canoncialvariate can explain the over 94.8% proportion of variance of data sets. So the first pair of canoncialcorrelation variate is representative. That shows the canoncial correlation analysis results is rathersatisfactory.Table 7 Canonical correlations and function testCorrelations Eigenvalue Difference Proportion Cumulative Pr1 0.9975203.5083 194.5966 0.9476 0.9476 <.0001 2 0.94828.9116 6.8612 0.0415 0.9891 <.0001 3 0.81992.0505 1.7534 0.0095 0.9986 0.0022 4 0.4786 0.2971 0.0014 1.0000 0.7710 0.3227Null hypothesis: remaining correlations are zero.The first pair of canoncial functions is constructed as follows:112345670.321 2.0440.0990.3340.260.470.1630.127U X X X X X X X =−+−−+−−−112341.4450.7920.461 2.099V Y Y Y Y =−+−+8X (3)From the standardized coefficients of the first pair of canoncial functions, it is found that themain factor of transportation is highway passenger traffic turnover volume (X 2). The main factor ofnational economy is GDP of tertiary industry (Y 4). The results show that development of tertiaryindustry is mainly influenced by the highway passenger traffic in 1978-2007.Table 8 Correlations Between the original y variables and the canonical variablesy1 y2 y3 y4V10.9757 0.973 0.9546 0.9909 U1 0.9733 0.9706 0.9523 0.9884Table 9 Correlations Between the original x variables and the canonical variables x1 x2 x3 x4 x5 x6 x7 x8U1 0.9335 0.9808 0.7125 0.9152 0.9292 0.9539 0.8894 0.9244V1 0.9312 0.9784 -0.7107 0.913 0.9269 0.9516 0.8872 0.9221Based on the results of Table 8 and 9, it is found that the U 1 is highly positively correlated with allX variables. V 1 is positively correlated with all Y variables. Also it is found that V 1 is positivelycorrelated with all X variables except X 3. U 1 is highly positively correlated with all Y variables. So thefirst pair of canoncial function can represent the original data sets effectively.4 ConclusionsFrom the above analysis, it is found that development of transportation industry are highly relevantto national economic growth. In planning economy period, the main factor of national economy is GDPof manufacturing industry (Y 2).The main factors of transportation is highway passenger traffic turnovervolume (X 2), railway freight traffic turnover volume (X 5). They are highly correlated. But in the reformand opening-up period of 1978-2007, it is found that the main factor of national economy is GDP oftertiary industry (Y 4). The main factor of transportation is highway passenger traffic turnover volume(X 2).The results are consistent with the fact that China has changed economic structure greatly frommanufacturing industry to tertiary industry. Also it is found that in planning economy the highwaypassenger traffic and railway freight traffic had played the most important role in national economic development. In the period of 1978-2007, highway passenger traffic had become the most important traffic mode in the development of tertiary industry. The highway traffic and railway traffic have contributed greatly to the national economic development. Therefore, governments at all levels and the whole society should implement active policies for the development of transportation industry espiecially for the highway industry and railway industry in order to achieve sustained economic growth in the current period..References[1]Xu Qingbin，Rong Chaohe,Mayun，etc.Introductionary transportation economics[M]. China’sRailway Press,1995.[2]Huang Zhendong. Relations between transport industries and national economy[J].Economics andManagement Research,1993(5), p1-3.[3]Rong Chaohe. Transportation[M]. China’s Social Science Press,1993.[4]Hanbiao. Development theory of transportation[M]. Dalian Marine University Press, 1994.[5]Xiong Rongjun,etc. Research on Transportation and economic development.Beijing JiaotongUniversity, Doctor thesis,1997.[6]Wang Chuanxu. Quantitative Study of Transportation’s Contribution to National EconomyGrowth[J]. China Journal of Highway and Transport，2004, 17(1),p94-97．[7]Wang Libin，Wu Qunqi.Discussion on contribution of highway investment to national economy[J].China Journal of Highway and Transport, 2006,19(3),p96-99.[8]Li Weidong. Applied multivariate statistical analysis[M].Peking University Press,2008.Author in brief:Li Weidong, associate professor, mainly engaged in statistics, enterprise management. E-mail: nju_lwd@ .Canoncial Correlation Analysis on the Relationship BetweenChina's Transportation Industry and National EconomicDevelopment作者：LI Weidong作者单位：School of Economics & Management,Beijing Jiaotong University, P.R.China,100044引用本文格式：LI Weidong Canoncial Correlation Analysis on the Relationship Between China's Transportation Industry and National Economic Development[会议论文] 2009。
Social dimensions of syntactic variation-- the case of when clauses
Social dimensions of syntactic variation: the case of when clausesJenny CheshireIntroductionPeter Trudgill and Jack Chambers hoped that their 1991 volume on grammatical variation in English dialects would stimulate further research into the grammar of nonstandard dialects and the nature of grammatical variation in language (1991:295-6). Since the publication of that volume there has been a good deal more research on these topics, from theoretical linguists as well as sociolinguists, and even from a collaboration between the two (for example, Henry and Wilson 1998). Mainly, however, researchers have focused on language internal constraints on variation, and the social dimension of grammatical variation has remained less studied. It has been argued, in fact, that syntactic variation is conditioned less by social factors than by internal, cognitive and situational constraints (see, for example, Rydén 1991, Scherre and Naro 1992), and that syntactic variation may rarely, if ever, serve the function of distinguishing social groups in the way that 'classic' phonological and morphological variants do (Winford 1996: 188, Hudson 1996:45). One reason for this would be the infrequency of syntactic forms relative to phonological or morphological variants: since syntactic variants are less frequently heard, they are presumably less likely to become associated with a specific social group and to function as sociolinguistic indicators or markers. But until we have clear evidence of the social patterning of syntactic variants in different communities, or the lack of such patterning, we cannot draw any firm conclusions about the relationship between syntactic variation and the social world.Discovering the social dimension of syntactic variation is, of course, less straightforward than for phonological and morphological variation. There are several methodological and conceptual problems, some of which are mentioned by Trudgill and Chambers, as we will see. In this paper I focus on some of these problems, suggesting the kinds of methodological procedures that might help resolve them. I will use as illustrative material an analysis of one type of clause found in conversations between 12-16 year old working class adolescents in Reading, Berkshire, recorded in adventure playgrounds during a nine-month period of participant observation1. This analysis will show that syntactic variation can be intricately involved in the construction of social meaning, but that the involvement is of a different kind from that of phonological and morphosyntactic variation.Lone when clausesAn initial difficulty in the analysis of syntactic variation lies in deciding what to analyse, as grammatical variants often evade the conscious awareness of speakers and listeners. Chambers and Trudgill point out that they are less easily disassociated from the discoursethan phonological variants and are not readily subject to paralinguistic comments or observations. Furthermore,contextual cues, felicity conditions and implicatures disguise (or perhapscompensate for) unusual syntactic structures in a conversation, whether thosestructures are the result of performance factors (false starts, ellipses, blends) ordialect differences (op.cit.: 292).The clause type that I will discuss in this paper seems a clear example of an unusual syntactic structure that may be overlooked. I did not notice it until I had decided to focus specifically on an analysis of the clause structures used by the adolescent speakers. For this it was necessary to work closely with the transcripts of the playground conversations, identifying every instance of a non-canonical clause, including any that could be attributed to performance factors. One such non-canonical clause is an adverbial clause introduced by when, with no accompanying main clause. I will refer to these clauses as lone when clauses. Extract 1 provides one illustration: other examples are given later in the paper.Extract 1 (the boys are talking about one of their teachers, who was married to someone I knew)Nobby: yeah Miss Threadgold she ain't badRob:yeah she she went camping with usJenny:yes he told me she'd been camping→Nobby:when we went campingRob:she's a good laughJenny:is she?Nobby:yeah(DAVE -do you think this is OK - 'Miss Threadgold' is a pseudonym for Miss Trudgill -Peter's first wife, who was these boys' teacher! Do you think this is too cheeky?)There are 28 lone when clauses in this data set, in a corpus of 50,000 words - not many, but enough for them not to be attributable to performance factors. Interestingly, this figure represents a far higher proportion of unaccompanied when clauses than the literature would lead us to expect. Lone when clauses account for 25 per cent of all when clauses in the playground conversations: there are 105 when clauses in total, 77 of which have an accompanying main clause. By comparison, in Ford's corpus of adverbial clauses in American English (see Ford 1993) only 3 per cent of the total number of temporal adverbial clauses had no main clause (there were only 2, in 63 temporal clauses). Furthermore, in Ford's data temporal clauses were the least likely to occur with no main clause: conditional clauses were the most frequent, although even so there were only 8 of these (15 per cent), from the total of 52 conditional clauses. Similarly, Mondorf's analysis (2000) of adverbial clauses in the London-Lund corpus found only 6 per cent of adverbial clauses with no main clause (259, out of 4462 clauses); again, these were mainly conditional clauses. The temporal clauses in Ford's study were introduced with a range oftime adverbs, of which when just one (Mondorf did not analyse temporal clauses). The frequency of lone when clauses in the Reading data set, then, marks them as an unusual phenomenon that is worth further analysis.Explicatory lone when clausesThe next step in the analysis was to identify the discourse functions of the lone when clauses. Four of them were used to explain something that had been mentioned in the previous discourse. In extract 1, for example, the discussion is about the excursions that the playground leader, Sue, organises. Mandy has trouble finding the word she wants to describe the destination of the next excursion, and after unsuccessful attempts by others as well as by herself she finally explains what she is referring to with her when clause. Extract 1Terry:Sue's given lots of outings…first one was Chessington ZooJenny:was it good?Terry: yeahMandy:next one is the ludo thing ain't itPaul:not a ludoMandy:judoJenny:judo?→Mandy:not a judo! when you're on a water chute and you go right in the ...er water.. you goes right down into the waterJenny:like a water chute thing?Mandy:yesSimilarly, in extract 2 Rob explains how Britt (another playground leader) tries to control her mind, providing a time frame for a specific situation that illustrates what she does: Extract 2Rob:and Britt she's queer = = she's trying to learn to control her mind Nobby: = yeah =Rob:whatever that meansJenny:is she?Rob:[yeahNobby:[yeahJenny:oh how is she going to what is she doing to conNobby: I don't know→Rob: when you look at smoke and that you know fire =Jenny: = yeahNobby:she looks at a flame she's. you can look at . she's trying to look at a flame until it burns right outJenny:and then w. how does that control your mind?Rob:I don’t knowThis interactional function of the lone when clauses corresponds in some respects to a function that has been previously noted for initial adverbial temporal clauses. Ford (1993:29, 32), for example, notes that initial when clauses can explicate a semantically broad term such as thing or then. In her data, however, the explication occurs within an extended speaker turn; she argues, in fact, that the use of the semantically broad terms contributes to the projection of extended turns. In the Reading data, the four lone when clauses are explicatory but they do not project an extended turn. Instead, they clarify a term that the emerging discourse shows to be ambiguous, or too vague for present purposes, or otherwise problematic. They occur in response to an question from another speaker, and the turn in which they occur is typically short, as are the subsequent turns. It is not only the lone when clauses that are used in this way: initial when clauses sometimes serve this purpose too, as in Ford's corpus. Initial when clauses with this function, though, bear a superficial resemblance to lone when clauses, since they are separated from their main clause by intervening speaker turns. For example, in extract 3, the when clause explicates the ambiguity of Sharon's pronominal referent she.Extract 3Julie:oh don’t kick her Tina!Sharon:did she hurt?Jenny: who? that little one?→Sharon:no her …when she kicked youJenny:oh her..not very hardJulie: that's what she's always doing…kicking everyoneJulie had told her younger sister Tina (too late!) not to kick me. I had not heard her say this and as I had previously needed to safeguard the tape recorder from a small child who tripped as she was running past and fell on me, I thought Sharon's question did she hurt might refer to that event. Sharon's when she kicked you then makes it clear that her question was about Tracy's kick rather than the other event. Thus the sequence can be seen as did she hurt when she kicked you, with the when clause separated from the main clause by my intervening turn, in which I show that I had not understood who Sharon's she referred to. The function of the when clause is to explicate the ambiguity: Sharon clarifies the referent with her no her and then provides further clarification by indicating the specific time frame with the when clause.Extract 4 provides a further example. Here Nick corrects my assumption that the shop he had previously mentioned was in Reading, adding when I went down town to what can be analysed as the main clause it was in his previous turn. The when clause gives precision, then, to the vague quite a long way.Extract 4Jenny: which shop was that?Nick:oh it was quite a long wayJenny:in Reading was it or…Nick: no when I went downtownPivotal lone when clausesThe remaining 22 lone when clauses have a different function: and unlike the explicatory lone when clauses they initiate an extended turn, usually a narrative of personal experience. There is an example of this in extract 5, from a conversation between two girls, Valerie and Christine, and one boy, Tommy:Extract 5Jenny:you have to do horrible jobs if you’re a nurse.. all the bed pansAll:<LAUGHTER>Jenny:have you ever been in hospital?Valerie:[I haveChristine:[oh yeah I haveValerie:I got run over by a carChristine:I fell off a gate backwards <LAUGHS> and I was unconscious→Tommy:oi when I.. when I went in hospital just for a little while…Valerie: sshhh Tommy:cos my sister and my cousin they bent my arm ..they twisted it right round Here a discussion about nursing as a possible career prompted me to ask whether they had ever been in hospital. Valerie and Christine each take brief turns to mention one occasion when they were taken to hospital. Tommy also mentions an occasion when he went to hospital, but he prefaces this with a lone when clause (also with an attention-getting oi!). This is interpreted by the other speakers as an indication that he intends to take an extended turn, as we see from the fact that Valerie compliantly tells her younger sister to be quiet; and Tommy goes on to tell the story of his stay in hospital.Lone when clauses were also used to initiate and mark a sequence of 'joint remembering' (see Edwards and Middleton 1986), where two or more friends go over scenes from a favourite TV show or film that they have all seen. Here they do sometimes project an extended turn, in which the speaker talks about a specific part of the film; often, however, they initiate a stretch of co-constructed talk where the speakers together relive their enjoyment of the event. Extract 6 illustrates this: it is part of a long sequence where four friends are discussing The Hunchback of Notre Dame, a film that had been shown on television recently (this is discussed in more detail in Cheshire 1999). The short extract contains three lone when clauses, all of which mention specific sections of the film that the speaker has enjoyed. The first nomination, from Johnny, is successful, as thefollowing turns show. Darren's bid is not at first successful: he stumbles momentarily, and the floor is immediately taken by Patsy, who proposed a different topic. Her topic is not taken up by the other speakers, however – Nicky, in his turn, shows that he does not think much of this episode – and Darren seizes the opportunity to continue recounting his favourite scene, adding detail to increase speaker and hearer involvement and to justify the telling.Extract 6Patsy: I thought he was gonna fall when he was treading on the em er edge→Johnny: when he was on them bloody bells swinging about and he Patsy: yeahknocked..and he knocked his master down didn't he from thegalleries…woo crash!→Darren: when he was gonna em→Patsy when he told that girl he was deaf .. he got deaf by the ...that bell = = I knowNicky: = but he was still there =Jenny: was that on the telly?All: yeahDarren: I saw it in colour..you could really see the bloodPatsy:[Darren…don't put me offNicky:[I was gonna see it againDarren:and I see this blood went slu-u-urp <DRAWLS> all come out .. I see it all come out of his mouthNicky:it was horribleAgain, this function of the lone when clauses corresponds to an interactional function noted by Ford for initial adverbial clauses in initial position: they form pivotal points in the development of talk and present explicit background for material that follows (1993:62). Like initial temporal clauses in Ford's corpus, they project an extended speaker turn. Initial when clauses can also have this function in the Reading conversations, as illustrated by extract d. Thus, as with the explicatory lone when clauses, the function of what I will term pivotal lone when clauses is shared by conventional adverbial when clauses.It is interesting that the lone when clauses in the Reading conversations correspond in their functions to what other researchers have noted for initial temporal clauses, but the fact that there are so many unaccompanied when clauses in the Reading conversations is puzzling. And although other researchers have assumed that it is possible to infer a main clause from the context for unaccompanied adverbial clauses, it is not usually possible to unambigously do so here (what, for example, could be inferred as the main clause for the when clauses in extracts 5 and 6?).A simple breakdown in terms of the gender of the speakers who use lone when clauses points to a further unexpected phenomenon: as Table 1 shows, pivotal lone when clausesare used far more frequently by male speakers than female speakers. There is no gender difference, however, in the use of explicatory lone when clauses.Table 1. Lone when clauses used by male and female speakers in the playgroundconversations______________________________________________________________________male speakers female speakerspercentage tokens percentage tokens total number of tokens pivotal lone when clauses91.7 228.3 224explicatory lone when clauses * 3 * 1 4total * 25 * 3 2_______________________________________________________________________________________________ * Numbers of tokens are too low for a percentage to be viableI then listened to all the lone when clauses in the audio recordings: this revealed that pivotal lone when clauses are distinguished from explicatory lone when clauses, and indeed, from all other when clauses, by a characteristic intonation contour. They are distinct in having level tones on every syllable except the last: this has a falling tone, and is slightly drawled.Pivotal lone when clauses, then, differ in form from the other when clauses in the data set, both in their lack of a main clause and in their characteristic intonation; yet their interactional function appears to be no different from that of conventional initial when clauses (and of initial temporal clauses introduced by other temporal conjunctions; see Ford 1993). They seem, then, to be candidates for a variationist analysis. We could set up a sociolinguistic variable for which pivotal lone when clauses would be one variant, and another variant would be when clauses with an accompanying main clause and the same interactional function. Pivotal lone when clauses could then be seen as sociolinguistic indicators, used predominantly by male speakers in these groups of friends. If we assume that the variants with a main clause are 'standard' forms and the pivotal lone when clauses are nonstandard (because they lack a main clause), we could interpret the variation as a manifestation of the well-attested sociolinguistic gender pattern, where female speakers use more 'standard' forms. This would also confirm the view that female speakers tend to delete less material than male speakers, and so are less redundant, more formal and more polite (Shibamoto 19 ). Syntactic variation, we could then say, serves to distinguish social groups in the same way as classic phonological and morphological variables.This approach, however, ignores two major problems associated with the analysis of syntactic variation. First, it does not address the question of what is 'standard' in spoken English apart from what is prescriptively defined in this way (Milroy and Gordon 2002: 185). Prescriptivists have nothing to say about adverbial clauses without a main clause, perhaps because they are good examples of forms that occur infrequently and so are not noticed (again, see Chambers and Trudgill 1991: 292), or perhaps because they occur only in specific settings or with specific groups of speakers. Second, seeing lone when clauses as an alternant to conventional when clauses is not necessarily the most appropriate way to conceptualise the variation in which the lone when clauses areinvolved. In this analysis I have been considering lone when clauses to be adverbial clauses, in line with both Ford's and Mondorf's analysis of clauses with no accompanying main clause. . But this relies on conventional frameworks of grammatical analysis, and these frameworks are heavily influenced by the form of written language. They may not, therefore, always provide the most appropriate categories for the forms that occur in spoken language. The functions of the lone when clauses overlap with some that have been reported for initial adverbial clauses, as we have seen, but adverbial clauses are, by definition, part of a main clause (Biber et al 1999: 194, Quirk et al 1985:1047). Lone when clauses can be considered part of a main clause only if we assume that a main clause is ellipted, but we have already seen that it would require a large stretch of the imagination to see many of the lone when clauses in this way.There are no other grammatical frameworks that point to alternative ways of conceptualising the variation. I had thought that Biber et al's (1999) grammar would be helpful, since this is based partly on spoken English: indeed, unlike other reference grammars it does include clause structures common in conversation. One such structure is an unembedded dependent clause (op.cit.: 223-4). However, the lone when clauses discussed here do not fit the description of this type of clause. The most important types, according to Biber et al, do not occur in turn initial position: instead, they allow speakers to add explanation or justification to what has already been said or to express a comment about what has been said. The lone when clauses occur in turn initial position, are structurally independent, and their interactional function, as we have seen, is to project an extended turn or propose a topic for joint remembering: the when simply points to the event that is proposed as the topic.A way forward is to look not merely at the form and the interactional function of the lone when clause but also at the content of the clause, to see what kind of events are pointed to by the when clause. We have already seen that sometimes the event is part of a film or TV show that the speakers have all seen and enjoyed: in the extended turn that follows the speakers, through their talk, relive their enjoyment. Looking at the narratives of personal experience that are introduced by a lone when clause showed that here too the narrative can often be seen to be about an event that is familiar to the other speakers, either because they took part in the events being narrated, or because the story is one that they have heard before. This is clear from the beginning of the narrative in extract 7, where the when clause is uttered with the characteristic intonation of the pivotal when clauses (with a falling tone on Wight) and the you know indicates that this is a familiar story (as does Alec's response).Extract 7Jeff:when we went to the Isle of Wight though [you knowAlec: [fucking hell he fell in loads of stinging nettles and the way he cried mate "oh God oh God oh God" Jeff: yesNarratives of this kind too, therefore, are a form of joint reminiscing. As other researchers have observed, retelling familiar stories is a way of reinforcing group membership, allowing participants to relive common experiences and confirming a shared long-term bond (Norrick 1997: 211).The question that needs to be answered now, then, is why male adolescents should use lone when clauses in this way while female speakers do not. In other words, does the lone when clause function simply as a sociolinguistic indicator, indexing the fact that the speaker is male, or does it indicate that male speakers in the adventure playgrounds propose familiar topics for their narratives or initiate sequences of joint remembering more frequently than female speakers do? If so, what are the male speakers accomplishing through this aspect of their talk? Further, is this something that the female speakers do not accomplish, or do they accomplish it using different linguistic means? These questions take us a long way from the analysis of syntactic variation, for it now becomes necessary to explore the narrative sequences that occur in the conversations. It is necessary, however, to follow this trail in order to discover why the gender difference might exist. SKIRTS COULD BE MENTIONED HERE…..The narrative analysis that I carried out is reported in Cheshire (2000). I will mention the main conclusions only briefly, in order to focus on the role of the linguistic variation in which the lone when clauses are involved. Analysing all the narratives in the 50,000 word corpus would have taken more time than was available, so it was necessary to be selective. The playground conversations are varied in nature: some have more narratives than others, some include more speakers than others. However, several of the lone when clauses used to introduce a narrative or reminiscence came from conversations where there were several boys talking to each other, so I chose for detailed analysis the three all-male conversations with the highest number of narratives and, similarly, the three all-female conversations where the most narratives were told. I isolated all the narrative sections in these conversations, using Labov's definition of a minimal narrative as consisting of at least two temporally ordered clauses following the order in which the real world events could be inferred to have taken place (Labov 1972: 360-61). Very few of the narratives in the conversations were as short as just two clauses, however. This gave a total of 124 narratives – 58 from female speakers and 66 from male speakers.The analysis revealed that a consistent characteristic of the narratives told in the male friendship groups was a concern to create a sense of group identity through the telling of a story. This was shown especially by central members of the friendship groups. In their monologues they explicitly marked their stories as familiar to the other speakers by addressing them by name in tags (for example, Nobby says I was pushing my granny and I bit my tongue in half didn’t I Ben), and they encouraged their friends to tell stories of their own about events that were known to them all. A high proportion of the boys' narratives was co-constructed, though often the nature of the co-narration prevented acoherent tale from being told. This was because many of the contributions from individual speakers were insults, contradictions, interruptions and other attempts to seize the floor. The pace of speech was fast, and all the speakers seemed to be enjoying themselves: it seemed to be less important for speakers to secure the floor and recount their story than to participate in group talk and the camaraderie it produces. The point of the contradictions was to show familiarity with the events being recounted rather than to challenge the speaker, as could be seen from the speakers' responses.Table 2 shows that 10 of the story openers used in the male conversations marked the story as a shared event about which they could jointly reminisce. Lone when clauses accounted for 5 of these story openers. In terms of their interactional function, then, these clauses can now be seen as one of a group of forms that propose a topic for a sequence of group talk. This group of story openers is used almost exclusively by the male speakers; there is just one token in the female narratives2. We gain a better understanding of why speakers use lone when clauses in their talk if we see the variation in which these clauses are involved in this way than if we think in terms of the form of adverbial clauses.Table 2 Story openers in the corpus of narrativesAll-female conversations All-male conversations Total Markers of a shared reminiscenceRemember when clause011 What about that time when022 You know when clause011What about X101 Pivotal lone when-clause055I can’t forget that time when011__________________________________________________________Total11011 Temporal subordinate clauses introduced by:once606 when527the other day505one time202one day303last time101yesterday101_____________________________________________________________________ Total23225 Zero opener93443 Miscellaneousclause right17 8left dislocation9312 there was X63 9you know X22 4you see30 3see20 1you should have seen X mate01 1 fuck me10 1oh yes01 1 fuck me01 1 he’s a bastard mate01 1oh it’s horrible10 1it wasn’t half fun10 1________________________________________________________________________ Total all story openers5864123The style of telling in the female conversations was very different. Monologues predominated, and other speakers were rarely (only once, in fact) drawn into the telling with a tag explicitly addressed to them. When narratives were co-constructed they consisted of a series of contributions from individual speakers, with one girl's contribution building on the previous speaker's, and sometimes expanding on it. Contradictions and interruptions were rare, and when another speaker added to the narrative it was usually to contribute points of details that the main speaker had explicitly asked for. In all the female narratives, the speaking rights of the current speaker were respected and a coherent tale was told. Their style of telling was, again, reflected in their choice of story openers: as Table 2 shows, 23 of these were temporal adverbial clauses, forms that were used almost exclusively by the female speakers. The when clauses used as story openers were conventional adverbial clauses dependent on a main clause: like the other temporal adverbial clauses used, their interactional function was to secure the floor by projecting an extended turn, whilst situating the event reported in the main clause within a specific time frame. Extract 8 provides an illustration. Here Julie answers my question about frightening events by beginning a story about a specific time when she and Valerie had been playing near the motorway (the M4) and had been frightened by a man who was following them. She begins her story with a conventional when clause, dependent on the main clause there was a man following us; her weren't h e asks for confirmation, which her friend tries to provide, and together they establish the background information to the story. Once this has been done Julie takes the floor and continues with her story.Extract 8Julie:if ever anybody says to me your mum told me to c.c. bring you home in my car and if I didn’t know this person I would say I would run.. or else Iwould knock at the nearest house。
The Measurement and Antecedents of Affective, Continuance and Normative Commitment to the
2
Natalie J. Allen and John P. Meyer
Not surprisingly, confusion surrounding the conceptual distinctions is reflected in attempts to measure the construct. Indeed, relatively little attention has been given to the development of measures of commitment that conform closely to the researcher's particular conceptualization ofthe commitment construct. Our intention here, therefore, is threefold: (1) to delineate the distinctions between three of the more common conceptualizations of 'attitudinal' commitment,* (2) to develop measures of each, and (3) to demonstra;te that these measures are differentially linked to variables identified in the literature as antecedents of commitment. The third aim serves the dual purpose of providing evidence for the convergent and discriminant validity of the new measures and of providing a preliminary test of hypotheses concerning the development of commitment. The conceptualization and measurement of attitudinal commitment Although several conceptualizations of attitudinal commitment have appeared in the literature, each reflects one of three general themes: affective attachment, perceived costs and obligation (Meyer & Allen, 1987 a). Affective attachment The most prevalent approach to organizational commitment in the literature is one in which commitment is considered an affective or emotional attachment to the organization such that the strongly committed individual identifies with, is involved in, and enjoys membership in, the organization. This view was taken by Kanter (1968) who described 'cohesion commitment' as 'the attachment of an individual's fund of affectivity and emotion to the group' (p. 507) and by Buchanan (1974) who conceptualized commitment as a 'partisan, affective attachment to the goals and values ofthe organization, to one's role in relation to the goals and values, and to the organization for its own sake, apart from its purely instrumental worth' (p. 533). The affective attachment approach is perhaps best represented, however, by the work of Porter and his colleagues (Mowday, Steers & Porter, 1979; Porter, Crampon & Smith, 1976; Porter, Steers, Mowday & Boulian, 1974) who defined organizational commitment as 'the relative strength of an individual's identification with and involvement in a particular organization' (Mowday et al., 1979, p. 226). Porter and his colleagues developed the Organizational Commitment Questionnaire (OCQ) to measure the commitment construct (Mowday et al., 1979). This 15-item scale has been used extensively in research and has acceptable psychometric properties. A parallel measure developed in Great Britain for use with blue-collar workers has also been shown to be 'psychometrically adequate and stable' (Cook & Wall, 1980, p. 39). Although other measures of affective attachment have been developed for use in specific studies, they typically have not been subjected to rigorous psychometric evaluation.
英语六级作文关于人际关系的题
英语六级作文关于人际关系的题The Essence of Interpersonal Relationships.Interpersonal relationships are the fabric that binds society together, shaping our lives and experiences. They encompass a diverse range of interactions, from the most casual acquaintances to the deepest and most intimate bonds. Understanding the dynamics of these relationships iscrucial in navigating the complexities of modern life.At its core, an interpersonal relationship is built on communication. Effective communication is the lifeblood of any relationship, whether it be a friendship, a family tie, or a professional partnership. It is through communication that we express our thoughts, feelings, and needs, and itis through listening that we understand and connect with others. The art of communication lies in finding a balance between being assertive and respectful, honest and compassionate.Trust is another essential component of interpersonal relationships. Trust is built over time through consistent, reliable, and honest behavior. When trust is established,it creates a safe space where individuals can be vulnerable, share their weaknesses, and rely on each other. Trustfosters a sense of security and stability, which is crucial for the growth and development of any relationship.Respect is another cornerstone of healthy interpersonal relationships. Respect is the acknowledgment of another person's dignity, worth, and rights. It involves treating others as equals, valuing their opinions, and respecting their boundaries. When respect is present, it creates a positive and supportive environment where individuals can flourish and grow.However, interpersonal relationships are not always smooth and harmonious. Conflicts and disagreements are inevitable, and it is how we handle these situations that determine the strength and durability of our relationships. Conflict resolution skills are crucial in managing these tensions. Active listening, empathetic understanding, andthe ability to communicate effectively are essential in finding common ground and resolving differences.Moreover, interpersonal relationships are dynamic and evolve over time. They require constant nurturing and maintenance to thrive. Regular communication, shared experiences, and mutual support are essential in keeping relationships vibrant and alive. Additionally,relationships need to be flexible, adapting to the changesin individuals' lives and the evolving nature of their interactions.In conclusion, interpersonal relationships are the glue that binds us together as a society. They are complex and multifaceted, encompassing a range of emotions, expectations, and dynamics. Understanding and appreciating the essence of these relationships is key in fostering healthy, meaningful connections with others. By practicing effective communication, fostering trust, respecting others, resolving conflicts constructively, and continuously nurturing our relationships, we can create a more connected and compassionate world.。
rel字典中的意思
rel字典中的意思英文回答：The rel attribute in HTML is used to specify the relationship between the current document and the linked document. It is used in the <link> tag, and it can take on a variety of values, each of which specifies a different relationship.Some of the most common values for the rel attribute include:alternate: This value indicates that the linked document is an alternate version of the current document. For example, you might use this value to link to a mobile version of your website.author: This value indicates that the linked document is the author's homepage.canonical: This value indicates that the linked document is the canonical version of the current document. This is useful for preventing duplicate content issues.help: This value indicates that the linked document provides help information for the current document.icon: This value indicates that the linked document is an icon for the current document.license: This value indicates that the linked document contains the license information for the current document.next: This value indicates that the linked document is the next document in a series.prev: This value indicates that the linked document is the previous document in a series.stylesheet: This value indicates that the linked document is a stylesheet for the current document.The rel attribute is a powerful tool that can be usedto specify the relationship between different documents. It is important to use the correct value for the rel attribute, as this can affect how search engines and other webbrowsers interpret your website.中文回答：rel 属性在 HTML 中用于指定当前文档与链接文档之间的关系。
对比关系的英文作文高中
对比关系的英文作文高中英文，Comparative Relationships。
Comparison is a fundamental aspect of human cognition, shaping how we understand the world and make decisions. In this essay, I will explore the nuances of comparative relationships in various contexts, drawing examples from everyday life and literature.Firstly, let's consider the realm of education. When I was in high school, I often compared my academic performance with that of my classmates. For instance, if I scored lower than my friend on a test, I might feel motivated to study harder to improve my grades. This kind of comparison can be both beneficial, as it drives us to excel, and detrimental, as it may lead to unnecessary stress or competitiveness.In literature, authors frequently use comparison to enhance their storytelling. In Shakespeare's play "Romeoand Juliet," the relationship between the two lovers is contrasted with the feud between their families, the Montagues and Capulets. This comparison highlights the theme of love versus hate, adding depth to the narrative and resonating with readers on an emotional level.Moving on to the realm of work and career, comparative relationships play a significant role in our professional lives. For example, when seeking a job, we often compare the salary, benefits, and work culture of different companies before making a decision. Similarly, within a company, employees may compare their performance evaluations or job responsibilities with those of their colleagues, influencing their job satisfaction and career progression.Outside of academics and work, comparativerelationships also shape our personal lives and social interactions. In friendships, we may compare our interests, values, and experiences with those of our friends to strengthen our bonds or navigate conflicts. For instance, if one friend is more outgoing while another is moreintroverted, understanding these differences can lead to more harmonious relationships.In romantic relationships, comparisons can beparticularly complex. People may compare their current partner with past relationships or idealized notions oflove portrayed in media. This can create both positive moments of appreciation and negative moments of doubt or dissatisfaction. Learning to navigate these comparisonswith honesty and empathy is essential for maintaining healthy relationships.Turning to cultural comparisons, globalization has made it easier for us to encounter and appreciate different cultures. When I traveled to China, I was struck by the cultural differences in communication styles, food preferences, and social customs compared to my home country. These comparisons enriched my understanding of diversityand fostered cultural appreciation.In conclusion, comparative relationships permeate every aspect of our lives, influencing how we learn, work,socialize, and perceive the world. By recognizing the power of comparison and approaching it with mindfulness and understanding, we can harness its benefits while mitigating its potential pitfalls.中文，对比关系的探讨。
会计和银行的关系英文作文
会计和银行的关系英文作文英文：Accounting and banking are closely related in the world of finance. As an accountant, I often work closely with banks to manage the financial affairs of my clients. Banks play a crucial role in the financial ecosystem, providing various services such as loans, mortgages, and investment opportunities.One of the key relationships between accounting and banking is the management of financial transactions. When I prepare financial statements for my clients, I often need to reconcile their bank accounts to ensure that all transactions are accurately recorded. This involves working closely with the client's bank to obtain the necessary information and documentation.Another important aspect of the relationship between accounting and banking is the role of banks in providingfinancing for businesses. When my clients are in need of capital to expand their operations or invest in new projects, I often work with banks to help them secure the necessary funding. This may involve preparing financial forecasts and business plans to present to the bank, aswell as negotiating the terms of the loan or line of credit.Furthermore, banks also rely on accountants to provide them with accurate and reliable financial information. When a business applies for a loan or seeks to establish a lineof credit, the bank will often require financial statements prepared by a certified public accountant. This helps the bank assess the creditworthiness of the business and make informed lending decisions.In addition, the relationship between accounting and banking extends to the regulatory and compliance aspects of the financial industry. Both accountants and banks are subject to various regulations and reporting requirements, and they must work together to ensure compliance with these rules. For example, banks rely on accountants to help them adhere to financial reporting standards and tax regulations,while accountants depend on banks to provide accurate and timely information for their clients.Overall, the relationship between accounting and banking is a symbiotic one, with each relying on the other to effectively manage the financial affairs of businesses and individuals.中文：会计和银行在金融领域密切相关。
并存关系的作文题目
并存关系的作文题目英文回答：In my opinion, the concept of coexistence is crucial in today's world. It is important for people from different backgrounds, cultures, and beliefs to live together peacefully and harmoniously. This idea of coexistence can be seen in various aspects of life, such as in relationships, communities, and even on a global scale.For example, in a romantic relationship, it isessential for both partners to respect each other's differences and find a way to coexist peacefully. This could mean compromising on certain issues, communicating effectively, and being open-minded. By practicing coexistence in a relationship, both individuals can grow and learn from each other, ultimately strengthening their bond.Similarly, in a community setting, coexistence isnecessary for people of different races, religions, and social backgrounds to live together in harmony. This could involve celebrating each other's traditions and customs, supporting one another in times of need, and working together towards common goals. By embracing diversity and practicing tolerance, a community can thrive and prosper.On a global scale, the concept of coexistence is more important than ever. With the rise of globalization and interconnectedness, it is crucial for countries to coexist peacefully and work together to address common challenges such as climate change, poverty, and conflict. By fostering mutual understanding, cooperation, and respect, nations can build a more sustainable and peaceful world for future generations.Overall, coexistence is not just a passive acceptance of differences, but an active effort to live in harmony with others. It requires empathy, patience, and a willingness to learn from one another. By embracing the idea of coexistence, we can create a more inclusive and compassionate society for all.中文回答：在我看来，并存的概念在当今世界至关重要。
二天思辨作文两者的关系
二天思辨作文两者的关系英文回答：The relationship between two days of contemplation is a complex one. On one hand, they are similar in the sensethat they both involve deep thinking and reflection. However, on the other hand, they also differ in terms of their purpose and approach.Contemplation is a process of introspection and self-reflection. It involves pondering over one's thoughts, emotions, and experiences in order to gain a deeper understanding of oneself and the world around them. Both types of contemplation, whether it is for one day or two days, require a certain level of focus and concentration.However, the purpose of the two days of contemplation may differ. For example, one day of contemplation may be focused on finding a solution to a specific problem or making an important decision. In this case, the individualmay spend the day thinking about different options,weighing the pros and cons, and considering the potential outcomes. On the other hand, two days of contemplation may be more open-ended and allow for a broader exploration of one's thoughts and emotions.In terms of approach, one day of contemplation may be more structured and goal-oriented. The individual may set specific objectives for the day and follow a systematic process to achieve them. They may use techniques such as journaling, meditation, or engaging in meaningful conversations with others to facilitate their contemplation. On the other hand, two days of contemplation may be more flexible and spontaneous. The individual may allow their thoughts to wander and explore different avenues ofthinking without any specific agenda or timeline.In conclusion, the relationship between one day and two days of contemplation is both similar and different. They both involve deep thinking and reflection, but they maydiffer in terms of their purpose and approach. Whether itis one day or two days, contemplation is a valuablepractice that can lead to personal growth and self-discovery.中文回答：两天的思辨之间的关系是复杂的。
对比关系英文作文模板
对比关系英文作文模板英文：When it comes to comparing and contrasting, there are a few key phrases that can be used to show the relationship between the two things being compared. Some common phrases for comparing include "similarly," "likewise," and "in the same way." These phrases are useful when you want to draw attention to the similarities between two things. For example, you might say "Similarly, both cats and dogs are popular pets." This makes it clear that you are comparing the two animals and highlighting their similarities.On the other hand, when you want to contrast two things, you can use phrases like "however," "on the other hand,"and "in contrast." These phrases are useful when you wantto draw attention to the differences between two things.For example, you might say "However, while cats are knownfor their independence, dogs are often seen as more loyal and affectionate." This makes it clear that you arecontrasting the two animals and highlighting their differences.中文：当涉及到比较和对比时，有一些关键短语可以用来显示两个被比较的事物之间的关系。
摄影与我们之间的关系英语范文
摄影与我们之间的关系英语范文The Relationship Between Photography and Ourselves.Photography, a medium that captures moments in time,has a profound relationship with each of us. It is not just a hobby or a profession; it's a way of life, a means of expression, and a tool for understanding the world. In this article, we delve into the intricate connections between photography and ourselves, exploring how it shapes our perspectives, influences our emotions, and even defines our identities.A Window to the World.Photography offers us a unique perspective on the world. Through the lens, we see the familiar with new eyes, discovering beauty, sadness, joy, and profound insights in the most ordinary of scenes. It allows us to pause, reflect, and appreciate the little moments that often go unnoticedin our fast-paced lives. Whether it's the soft rays ofsunshine filtering through the leaves of a tree or the emotional exchange between two loved ones, photography captures these fleeting moments and immortalizes them for eternity.A Means of Expression.Photography is a powerful form of self-expression. It gives us the ability to communicate our thoughts, feelings, and visions without the need for words. Through the composition of images, we can convey complex emotions and stories that resonate deeply with the viewer. Whether it's through landscape photography, portraiture, or street photography, each click of the shutter is a personal narrative, a snippet of our lives and perspectives.A Tool for Understanding.Beyond capturing moments and expressing ourselves, photography also serves as a tool for understanding. It allows us to delve into the lives and stories of others, to see the world from their perspective. Through the lens, wecan explore cultures, histories, and environments that are different from our own, broadening our horizons and challenging our preconceptions. This understanding, in turn, helps us to be more empathetic, compassionate, and tolerant of the diverse world we live in.A Pathway to Creativity.Photography encourages creativity and experimentation.It challenges us to see the world in new ways, toexperiment with lighting, angles, and composition to create unique and captivating images. This creative process notonly enhances our skills but also ignites our imaginations, sparking a sense of wonder and curiosity that keeps us engaged and excited about the art of photography.A Means of Connection.Photography has the unique ability to connect us with others. It can be a powerful medium for storytelling, allowing us to share our experiences, memories, and perspectives with the world. Whether it's through socialmedia, exhibitions, or simply sharing photos with friends and family, photography breaks down barriers and brings people together, creating a sense of community and belonging.In conclusion, photography is not just a hobby or a profession; it's an integral part of our lives. It shapes our perspectives, influences our emotions, and defines our identities. Through the lens, we see the world in new ways, express ourselves deeply, understand others better, and connect with the world in meaningful ways. As we continue to explore the world through the eyes of our cameras, we discover more about ourselves and the beauty that lies within us all.。
月亮与六便士的关系英语作文
The Relationship Between the Moon and the SixpenceThe moon and the sixpence, though seemingly unrelated, hold a profound symbolic relationship that often finds its way into literature, art, and the human imagination. This relationship explores the contrast between the idealistic, ethereal realm of dreams and aspirations, and the practical, materialistic world of daily life.The moon represents the dreamer's ideal, the far-off, yet alluring goal that one strives for. It is the symbol of the unknown, the unexplored, and the mysterious. The moon's ethereal beauty captivates the imagination, inspiring us to aspire to higher things, to pursue our dreams, and to reach for the stars.The sixpence, on the other hand, is a tangible representation of the material world. It is a symbol of wealth, stability, and security. The sixpence represents the daily grind, the mundane tasks and responsibilities that occupy our time and energy. It is the practical side of life, the reality that we must confront and deal with on a daily basis.The relationship between the moon and the sixpence is a delicate balance. On one hand, we must aspire to our dreams and strive forsomething greater than ourselves. The moon calls us to leave our comfort zones, to challenge ourselves, and to seek new horizons. However, we cannot neglect the material world entirely. The sixpence reminds us that we need to earn a living, provide for ourselves and our families, and fulfill our responsibilities.The key is to find a balance between the two. We should not be so focused on the material world that we forget to dream and aspire. At the same time, we should not be so caught up in our dreams that we neglect our responsibilities and obligations. The moon and the sixpence together represent the perfect balance between the idealistic and the practical, the dreamer and the realist.In conclusion, the relationship between the moon and the sixpence is a metaphorical representation of the balance we must strike between our dreams and aspirations and the practical realities of daily life. It is a reminder that we should always strive for something greater, while remaining grounded in the world we live in.。
Classical Prescriptions
illUStRated tcmClassical PrescriptionsClassical prescriptions are those that have been proven by doc-tors down through the ages to have remarkable curative effects.Now more than 100,000 prescriptions fall into this category.In traditional Chinese medicine there are eight principals used to guide diagnosis, namely yin and yang , exterior and interior, cold and heat, and deficient and excessive symptoms.Classical prescriptions are an important part of traditionalChinese medicine. Their origin dates back to the time of legendary Chinese ruler Shennong over 5,000 years ago, and they were first recorded in the Shennong’s Herbal Classic , Canonical Methods for Brews and Decoctions , and Treatise on Febrile and Miscellaneous Diseases.The Principle of Compatibility of ingredients in Classical PrescriptionsThe principle of compatibility is described as the relationship between “monarch, minister, assistant and guide,” signifying dif-ferent roles played by different ingredients of a prescription in their actions. Rational combination of ingredients makes the prescription more effective.Source: Shandong Tongda Animation Games Production Co., Ltd.In the 2002-3 outbreak, Classical prescriptions were used to successfully fight against SARS in thousands of cases. The World Health Organization confirmed that traditional Chinese medicineprescriptions have good results alleviating the symptoms of SARSand reducing side effects caused by the use of hormones.This prescription, invented by Han Dynasty doctor Zhang Zhongjing (circa 150-219), consists of cinnamon twig, peony,licorice root, date and ginger. It is used to treat headache and fever caused by the common cold.Made from powders of peony, balloonflower root, and unripe bitter orange, this prescription is used to treat bronchial asthma, cough and phlegm retention.This prescription, consisting of pinellia tuber, scutellaria root, dried ginger, ginseng and licorice root, is from the Han Dynasty Treatise on Febrile Diseases . It is effective in treating gastritis, and is used by heavy drinkers to prevent stomach problems.。
正则动量面面观
正则动量面面观徐湛【摘要】对于带电粒子在磁场中的运动,一个很重要的概念是粒子的正则动量不同于它的机械动量,而这一点经常被忽视或者被混淆.本文从正则动量的基本定义出发,从经典的和量子的两方面分析了正则动量和机械动量的关系,指出正则动量才是基本的动力学变量.最后以磁镜装置为例说明了如何应用正则动量简明地给出问题的答案.%For movement of a charged particle in a magnetic field,a very important concept is that the canonical momentum of the particle is different from its mechanical momentum,which is often overlooked or confused.Based on the fundamental definition of canonical momentum,this paper analyzes the relationship between canonical momentum and mechanical momentum from classical and quantum aspects,and emphasizes that the basic dynamic variable is the canonical momentum.Finally,as an example,the magnetic mirror device is studied to illustrate how the canonical momentum concisely gives the answer to a problem.【期刊名称】《物理与工程》【年(卷),期】2017(027)005【总页数】7页(P3-9)【关键词】磁场;矢量势;正则动量;规范变换;磁镜装置【作者】徐湛【作者单位】清华大学物理系,北京 100084【正文语种】中文提起粒子的动量p这个概念，人们的第一反应通常就是p=m，但这并不总是对的。
关系英语作文
关系英语作文Title: The Importance of Relationships in English CompositionIntroduction:The art of writing an English composition is significantly enhanced by the exploration and portrayal of relationships. Relationships add depth, meaning, and relatability to essays, enabling readers to connect emotionally with the content. This document aims to emphasize the importance of incorporating relationships in English compositions.Body:1. Enhancing Engagement:Relationships between characters or concepts in an essay captivate readers, making the narrative more engaging. By developing intricate relationships, writers can create tension, conflict, or harmony, drawing readers into the story and keeping them invested in the outcome.2. Developing Characters:In narrative essays or stories, the portrayal of relationships helps in shaping characters. Interactions and connections between individuals reveal their personalities, motivations, and growth. This enables readers to understand characters on a deeper level, fostering empathy and emotional investment.3. Exploring Themes:Relationships serve as a powerful tool to explore and conveythemes in English compositions. Whether it is friendship, love, family, or societal relationships, these connections help writers delve into universal themes that resonate with readers. By doing so, the essay communicates valuable messages and prompts introspection.4. Adding Complexity:Incorporating relationships adds complexity to an English composition. The intricacies of human interactions create layers of meaning, allowing writers to explore various perspectives and emotions. This complexity keeps readers intrigued and encourages critical thinking.5. Improving Descriptive Language:Relationships provide a platform for writers to showcase their command over descriptive language. By depicting emotions, gestures, and dialogue, writers can paint a vivid picture of the relationship, enabling readers to visualize and feel the connection between characters or concepts.6. Fostering Relatability:Relationships are a fundamental aspect of human life, making them relatable to readers. When an English composition effectively captures the essence of relationships, readers can identify with the experiences, emotions, and dynamics portrayed. This relatability creates a sense of connection and engagement, making the essay more impactful.Conclusion:In conclusion, the inclusion of relationships in English compositions significantly enriches the writing experience. It enhances engagement, develops characters, explores themes, adds complexity, improves descriptive language, and fosters relatability. Writers should strive to incorporate relationships effectively to create compelling and meaningful essays that resonate with readers.。