This project is for the Kaggle competition "Predict Future Sales" https://www.kaggle.com/c/competitive-data-science-predict-future-sales.
The sales dataset includes daily historical data (time series data) from January 2013 to October 2015. We are asked to predict the total sales for every product and store in the next month i.e. November 2015.
There are three stages in my project:
- Define the problem
- Task: predict monthly sales
- Since we have time series data, it is natural to think of using the long short-term memeory (LSTM) recurrent neural network for prediction
- EDA
- Visualize the historical trends and seasonality
- Price trend
- Outlier analysis
- Item category analysis
- Model creation and training