Analyzing Urban Transportation Services using RideShare Data Insights

In this fast-growing world, the need for choosing or booking a ride is increasing. Since the fare displayed on booking apps such as Uber, Rapido, etc., and the fare after the ride is completed differ, it is imperative to predict the fare. There is a large difference in those prices. To address these issues, an optimized model for ride fare prediction should be there. These rideshare services brought a large change in the transportation system but predicting the fare system has to change. The study presents various machine-learning models of regression such as the Selection Operator regressor, Elastic-Net regressor, Ridge regressor, Random Forest regressor, Least Absolute Shrinkage, and XGBoost regressor. The XGBoost regressor outperformed its counterparts with an R2 value of 0.9734. After the regression models are applied the label values are classified into three categories: high, medium, and low. The obtained ordinal label values are given into classification algorithms like Naive Bayes, Extra Trees classifier, Support Vector Machine, and Multi-layer Perceptron are applied. A comparative analysis is done and SVM performed the best in classification algorithms with an F1 score of 99.53.