Buying Scenario and Recommendation of Purchase by Constraint Based Sequential Pattern Mining from Time Stamp Based Sequential Dataset

In today's competitive world, one of the most important drivers for sustainability and growth of a business is retention of customer. For retention of customer, it is important to understand customers’ current needs and wants. Understanding of customers’ buying behavior helps in designing products by businesses for ensuring product attributes that will result in customer satisfaction on purchase and consumption. Proposed work mainly focuses on prediction of few items that customer is in all probability more likely to purchase in future than other items based on data analysis done using sequential pattern mining technique. Proposed work additionally focuses on recommendations to be made for items of purchase in near and long term future, based on incorporation of Gap constraint in conventional FP-growth based PrefixSpan algorithm. Combination of Length, Gap and Item constraints help to recommend next highly probable items for purchase, based on increasing support and confidence values. Simulation and experimental study done on six IBM generated synthetic dataset which throw light on customers’ buying patterns in terms of purchases that happened regularly or happened just once. Average 45%-55% purchase happens only once which doesn’t lead to probability of customer making next purchase (IBM generated synthetic dataset). Proposed system focuses on next purchase item prediction based on support, confidence and gap constraint. Decision maker can analyze the reasons of current situation and accordingly make vital recommendations in terms of next products to be offered to customer for purchase, thereby increasing the probability of sales actually happening.