Forming a taxi service order price using neural networks with multi-parameter training

Abstract

Abstract The article discusses the use of neural networks to forecast the prices of orders in the taxi service. The analysis shows that the results of the price control action are preferable to the results that are obtained for the direct cost of the trip. Special attention is paid to such factors as the time of the order, the weather condition of the environment, the number of cars in the area, as well as the distance between the starting and ending points of the route. It was shown that networks ensure maximum accuracy with backward propagation of error with the number of neurons (5) close to the number of parameters (4).