Description
Forecasting is very critical to businesses. It helps them plan for the future and gives them an opportunity to be prepared for upcoming demand. It is often required in practice to estimate demand for thousands of products, e.g SKUs in retail stores. It s not humanly possible to create these many models. In this session, we show how we did this for a large retailer.
Outline of Talk
- Importance of Forecasting
- Overview of Time series Forecasting Techniques
- ARIMA
- Exponential Smoothing
- XGBoost
- Generating Prediction Intervals for the Forecast
- Our approach for forecasting large number of products
- Problem Statement
- Solution Methodology
- Code Walk through
- Questions and Answers