23 lines (18 loc) · 779 Bytes

NEW LOCATION FOR UPDATED CODE: https://github.com/mollyryanruby/auto_forecast

5 Machine Learning Techniques for Forecasting Sales

Objective:

Predict the number of monthly product sales using regressive and time-series modeling techniques. Paper: https://medium.com/towards-data-science/5-machine-learning-techniques-for-sales-forecasting-598e4984b109

Featured Techniques:

EDA
Linear Regression
Random Forest Regression
XGBoost
Long Short Term Memory (artifical recurrent neural network)
ARIMA Time Series Forecasting

Results:

Best results were obtained from the XGBoost and LSTM models
All models predicted within 2% of monthly mean sales for 12 month prediction

Data Source:

https://www.kaggle.com/c/demand-forecasting-kernels-only/data