离散选择模型logit模型实例stata分析.
合集下载
相关主题
- 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0 m1-1-5
m2-1-5
m3-1-5
m1-1-6
m2-1-6
m3-1-6
m1-1
m2-1
m3-1
Conclusion
• The best model is Model 1-1 • Statistical results of the estimation indicate that
THANK YOU
Discrete Choice Analysis Term paper
Logit modeling with stata
2016. 12. 14 G201538010 LI TINGTING
Contents
• Introduction • SP Data set information • Modeling scenarios setting • Data modification • Modeling estimated results • Modeling comparison • Conclusion
market segmentation makes the model more comprehensive. • The more variables model has, the more comprehensive model is.
❖ The model which is used 3 variables is more comprehensive than the model used 2 variables.
Model 4 is ignored
Data Modification
• Basic data
rail
truck
id distance question type
cost
time
los
cost
time
los
choice
223
5
1
210000
14
60 280000
6
60
2
223
5
2
260000
14
80 350000
MODEL 3-2 2variables (cost/LOS)
MODEL 4-2 2variables (time/LOS)
Data Modification
• We modify row-data to remove unreasonable data set
- Such as the choice of the not-dominant alternative
models.
SP Data set information
• Stated pointed: 2007 • Analysis commodity: steel • Analysis range: 30 people(row 600-1500) • Dependent variable: choice • Independent variables: cost [log(#/10^5)]
7
60
2
• Data reorganize example
id mode distance
qtype
cost
time
223 rail
5
1
210000
14
223 truck
5
223 rail
5
1
280000
6
10
210000
19
223 truck
5
10
280000
7
los
choice
logcost5
los2
time [hour] LOS [#/10]
Modeling Scenarios Setting
Modeling scenarios
Market Segment?
MODEL 1-1 - Distance 5/6
3variables (cost/time/LOS)
Y
MODEL 2-1 - Distance 5/6 2variables (cost/time)
✓ Model 1 has 1 unreasonable data sets(in all data sets) ✓ Model 2 has 31 unreasonable data sets(in all data sets) ✓ Model 3 has 8 unreasonable data sets(in all data sets) ✓ Model 4 has 85 unreasonable data sets(in all data sets)
MODEL 3-1 - Distance 5/6 2variables (cost/LOS)
MODEL 4-1 - Distance 5/6 2variables (time/LOS)
MODEL 1-2 N 3variables (cost/time/LOS)
MODEL 2-2 2variables (cost/time)
60
0
0.741937
6
60
1
1.029619
6
100
0
0.741937
10
60
1
1.029619
Baidu Nhomakorabea
6
Modeling Estimated Results(DIST5)
Model distance5
1-1-5
2-1-5
3-1-5
0.2899 0.2884 0.1042
Modeling Estimated Results(DIST6)
Introduction
• This paper developed a disaggregated logistics demand models using discrete choice analysis method.
• Data used is 2008-SP data from a survey. • Stata was employed for the estimation of logit
Model distance6
1-1-6
2-1-6
3-1-6
0.2588 0.2539 0.0704
Modeling Estimated Results(DIST5&6)
Model Distance5&6
1-2
2-2
3-2
0.2539 0.2521 0.0838
Modeling Comparison