Research on Short-term Traffic Demand of Taxi in Large Cities Based on BP Neural Network Algorithm

In this paper, a prediction model of urban taxi traffic demand based on BP neural network is proposed according to the characteristics of the increasing demand for public transportation. This model uses the method of zoning statistics to analyze the traffic demand of urban Taxis, and the traffic generation and traffic disappearance in a period are taken as the input of the network according to the statistical results. This paper extends the hidden layer to 2 layers and introduces the tanh activation function. Also, this paper selects passenger traffic data from 2016 to 2019 as the training set. After 1,000 epoch training, the training set Mean Square Error (MSE) is achieved 4.332 × 10
^–5
and the test set loss is achieved 0.01139.