Udacity Machine Learning Engineer Nanodegree Capstone Project
Predictive rider-driver demand-supply models allow DiDi Chuxing, China’s largest ride-hailing company, to direct drivers to where riders will need to be picked up. The Di-Tech Challenge, allows contestants to use real-data to build their own model, and earn a chance to win $100K in prize money and score a job at Didi’s Research Lab Beijing or Silicon Valley.
Dataset download: http://research.xiaojukeji.com/competition/main.action?competitionId=DiTech2016
The project structure assumes your dataset folder "season_1" or "season_2" is in your project root.
To run the experiment:
For benchmark model SVR simply run support_vector_regression.py
For initial model run train_predict() in decision_tree_regression.py
For final model simply run decision_tree_regression.py