Recommendation Engine for Retail Domain Using Machine Learning Techniques

A recommendation system is highly beneficial to any business in the retail industry; not only does it increase revenue but it also enables retailers to provide their customers with the products they require. In this paper we are building a recommendation engine for a retailer that analyzes available products data to make predictions of which products will gain more revenue to retailer and which products are from which supplier. This can be done by performing clustering of the products based on the transaction history of the product purchases and then dividing the products into low‐selling products, medium‐selling products and high‐selling products using canopy K‐means clustering technique. RFM analysis is performed to classify customers and suppliers. Recency Frequency Monetary technique is a Marketing strategy used for analyzing customer behavior such as how recently a customer has purchased (recency), how repeatedly the customer purchased (frequency) and how much customer spends (monetary). It also classifies the suppliers into Low Gain, High Gain and Medium Gain which depends upon the product transactions of each of the suppliers. Which products will gain more revenue in the future and from which supplier is determined, and how many must be purchased on an average is predicted using ARIMA model.