python 3.7
numpy, pandas, xgboost, lightgbm, sklearn, matplotlib, pickle
You can just run my lightgbm.ipynb step by step, finally you can obtain the predicting result, and the predicting result is written in the lgbresult.csv file. I used two datasets for training, one is train.csv from Kaggle, the other is feature.csv which contains some weather datasets. The trained model is saved in "lgb_reg2feature_best.dat"