• Perform regression models and tested on real data to find proper model without over-fitting

• Estimate complexity of learning such as generalization, estimation of error on new data to generate proper model

• Studied semi-supervised and unsupervised methods to classify different data

• Trained models for Corporacion Favorita Grocery Sales Forecasting at Kaggle

• Preprocess over 123,497,040 data sets which were used for feature selection and data training; Model training such as Xgboost, Lightgbm, Linear regression and Moving average are used in Python and package pandas.

• The LB on Kaggle is 0.727 with ranking 734/968