Analyzing game pattern in League of Legends with Machine Learning
- Based on League of Legends Ranked Games (Mitchell J / Kaggle 2017)
- EDA : bar, box, pie charts, correlation plots, additional data (champion info) into the original dataset
- Feature Engineering : created (tags) & dropped (bans, gameid, etc.) different features
- ML methods : Random Forest, KNN, SVM, GNB and XGB with hyperparameter tuning (ROC/AUC, precision, recall, f1, accuracy metric)
- Interpretation : PDP isolate plots, shap values and summary plot