Kaggle competition for prediction of future revenue per coustomer
This repo is only upto the first stage of the compititona and was in the top 9% in the standing leaderboard.
Dataset can be found at https://www.kaggle.com/c/ga-customer-revenue-prediction
This consists of Exploratory data analysis and model using LightGBM(Light Gradient Boosting Model) which has Faster training speed and higher efficiency and considerably Lower memory usage.
For more information about LightGBM you can acesses to https://lightgbm.readthedocs.io/en/latest/# I found many public kernals usefull in understaging the dataset.
The validation score is 1.5237. The public score on leaderborad is 1.435.