Predicting Total Fare of Taxi Rides

The Dynamics of Ride-Hailing Pricing

Predicting the total fare of a ride-hailing trip involves understanding various dynamic factors including distance, time of day, traffic conditions, and base rates. In the Data Science Championship by MachineHack, the objective was to accurately predict these fares.

Regression Techniques

I designed multiple regression models including CatBoost, XGBoost, LightGBM, and Multilayer Perceptron Regression. The ultimate solution involved an ensemble technique combining Decision Tree Regressor, Random Forest, XGBoost, and CatBoost.

Conclusion

By blending the predictions of these diverse tree-based models, the ensemble effectively minimized the mean squared error, capturing both linear relationships and complex non-linear interactions within the trip data.