Challenge

The biggest challenge was to develop a highly accurate model of demand forecasting for retailers to manage their inventory as retailer’s cost of inventory was rising significantly

Solution Approach

We used different advanced machine learning capabilities like LSTM/ XGBoost/ GBM, to find the best algorithm which could accurately forecast expected sales of an item considering its attributes, price, number of substitute products and their prices

Outcome

• More than 90% accuracy was achieved over a period of time