- This project is taken up from Kaggle, an online community of data scientists and machine learners (link to competition)
- The goal of the project is to predict PUBG player’s win place percentage based on the player’s statistics.
- Millions of players data is used for building a regression model.
- Players data within a group in a match is combined to reduce the dataset size.
- Some insightful derived features are added to improve the predictions.
- The best model using Light Gradient Method achieved a Kaggle leaderboard score of 0.0214 on test data.
- For more information about data analysis and various experiment results, kindly look into doc.pdf file.
- Final code for prediction is given in main.py file.
mcabinaya / kaggle-pubg-finish-placement-prediction Goto Github PK
View Code? Open in Web Editor NEWKaggle Competition - Analysis and prediction of PUBG players' finishing placement based on their final stats
Home Page: https://www.kaggle.com/c/pubg-finish-placement-prediction/data