The online ride-hailing service is a new travel mode based on mobile Internet and smartphones, which can provide passengers with the convenience of the last mile travel. Because the supply-demand relationship in different parts within the city is always changing, an accurate ride-hailing demand prediction can better help the platform to dispatch vehicles and match drivers with potential passengers, thus improving the efficiency of the whole urban traffic. In this paper, six machine learning methods are utilized and compared, for predicting the ride-hailing order number, including Linear Regression, Gradient Boosting, Random Forest, Support Vector Regression, XGBoost, and Multilayer Perceptron. The experimental results on a real-world dataset collected in Haikou, China by Didi Chuxing show that Random Forest is superior to other methods and achieves the smallest RMSE and MAE.
Paper
Full text
Ride-hailing Demand Prediction with Machine Learning
Semantic Scholar · Computer Science · 2022
Abstract
The online ride-hailing service is a new travel mode based on mobile Internet and smartphones, which can provide passengers with the convenience of the last mile travel. Because the supply-demand relationship in different parts within the city is always changing, an accurate ride-hailing demand prediction can better help the platform to dispatch vehicles and match drivers with potential passengers, thus improving the efficiency of the whole urban traffic. In this paper, six machine learning methods are utilized and compared, for predicting the ride-hailing order number, including Linear Regression, Gradient Boosting, Random Forest, Support Vector Regression, XGBoost, and Multilayer Perceptron. The experimental results on a real-world dataset collected in Haikou, China by Didi Chuxing show that Random Forest is superior to other methods and achieves the smallest RMSE and MAE.