Dynamic Pricing Prediction for Cabs Using Machine Learning

Abstract

Ride-on-demand (Rod) services like Uber and OLA cabs are becoming increasingly popular. Rod services employ dynamic pricing to strike a balance between supply and demand in an effort to improve service quality in order to assist both customers and drivers. However, dynamic prices frequently cause issues for passengers: They are unable to quickly make decisions due to frequently "unpredictable" prices. Therefore, it is essential to provide passengers with more information in order to address this issue, and a feasible solution is to forecast dynamic prices. In this paper, we focus on the estimation of dynamic prices by forecasting the price for each individual passenger order using the Rapido dataset as an example. Passengers will be able to alleviate their concerns by understanding, through price predictions, whether they will be able to obtain a lower price in the near future or in nearby locations. The prediction is carried out by learning the relationship between the dataset's features and dynamic prices. We train one linear model and test its output using real service data from various angles as a representative. In addition, on the basis of the model, we examine the contribution of features at various levels and determine which features are most responsible for dynamic prices. Finally, we use evaluation results to predict dynamic prices with an effective linear regression model. As accurate forecas, our hope is that the study contributes to passenger happiness