earticle

영어

This study deals with the taxi supply models based on the urban characteristics. The objective is to develop the appropriate taxi supply model to the urban characteristics by type, after classifying Korea cities and counties with the various urban indices. The study gives particular attentions to the multiple regression analyses and neural network models. Five significant factors are induced, through the factor analysis based on the 18 urban indices. All cities and counties are classified by 4 groups according to the cluster analysis. F-tests indicate that the differences of taxi supply size exist among urban groups classified by factor 1, 2 and 3. Multiple regression analyses show that the models by urban groups classified by the factor 1 are well fitted to the given data. Also the neural network models are better than multiple regression models in the view of fitness to the given data.