Campus travel mode prediction method integrating factor analysis and nested logit model

Abstract

The invention discloses a campus travel mode prediction method integrating factor analysis and nested Logit model, comprising: screening explanatory variables significantly correlated with travel mode selection and cycling distance level respectively based on the original survey data of non-motorized travelers; extracting common factors for travel mode selection and common factors for cycling distance level respectively based on the screened explanatory variables; constructing a two-layer nested Logit model with travel distance level as the upper selection branch and non-motorized travel mode as the lower selection branch; inputting the common factors for travel mode selection into the utility function of the lower selection branch, and inputting the common factors for cycling distance level into the utility function of the upper selection branch; calibrating the parameters of the two-layer nested Logit model using travel behavior data to obtain a travel mode selection probability calculation model; substituting the explanatory variable data of the traveler to be predicted into the probability calculation model to obtain the probability of each non-motorized travel mode.