Prediction for Public Transit Trip Frequency Using Ordinal Logistic Regression

The publics' transit trip frequency determines the demand for urban public transit directly. The accuracy of forecasting plays a key role in planning and improvement of public transport system. To build a prediction model of public transit trip frequency on the basis of limited data survey, taking the public transit trip frequency as the research object, selecting independent variables from the characteristics of residents, and the public transport service level. Predicting the model by using ordinal logistic regression and validating the model through the parallelism test, the Wald test and verification test. Taking Taiyuan as an example, 723 valid questionnaires have been collected by issue questionnaires in four locations. The results show that: three independent variables including residents' age, average monthly income, and travel time period are significantly related to the public transit trip frequency. The forecasting model is of statistical significance, and the forecasting accuracy rate is about 60.44%.