Inventory management using Demand Prediction
The client was facing a huge problem of stock pile-ups and stock outs for slow selling products and fast selling products respectively. This in turn caused huge losses to the client which resulted from the maintenance/storing costs of excess stock and the opportunity cost lost due to the insufficient stock. They wanted to design a dynamic inventory management system that can accurately forecast demand for all its products.
Python, SQL, Excel and Tableau (for visualization)
We provided a highly automated solution which made use of artificial intelligence. Our solution involved the development of various time series models which used different techniques such as LSTM, RNN (using sliding windows) and Logistic Regression. The final production model was an ensemble of these three models and the implementation of this at the client’s end resulted in a reduction of almost 32% in total costs.
Call: 020 3573 6917