Notebook: kaggle-bike-sharing.ipynb
The data set for this competition consists of
Since our task is to predict a numerical value (bike count) we employ a variety of regression algorithms. In particular, we explore:
1. Linear Regression
2. Ridge Regression
3. Lasso Regression
4. K-Nearest Neighbors
5. Decision Tree
6. Random Forest
7. AdaBoost
8. XGBoost
After tuning hyperparameters with GridsearchCV, our best model achieves a test score within the top 5% of the Kaggle leaderboard. For additional details and background information related to this dataset, see the Kaggle competition page at https://www.kaggle.com/c/bike-sharing-demand/overview.