An intelligent product suggestion algorithm using predictive analysis for personalized UI

Abstract

This research was done to introduce a personalized interface for item checkouts. Often we could observe lengthy queues in food outlets, supermarkets and other types of stores where the demand is high. Many studies are currently underway to find solutions to reduce the queue lengths and provide better service and satisfaction to the customers. This thesis reports a study of such scenario and how data analysis could provide a simple solution to the problem. As people expect more personalized experience nowadays, the solution for the problem was suggested as a personalized UI for the customers to select the items which they have purchased and create a checkout list all by himself/herself. To generate personalized content, an algorithm was implemented to output the next purchasing item set using the historical purchasing records of the users. The algorithm uses a rule based approach with weighted ratings. Although collaborative method is a popular method in finding such results, in the studied scenario, it is not applicable as the store does not maintain a comprehensive user profiles or facilitate the uses to rate products. The research introduce a model named RFR-U model. The model uses the parameters; relevance, recency and frequency to determine the next purchasing item set. Since the algorithm could generate the results with-in seconds it could be used in real time applications. During the research a dashboard and a mobile application was also implemented to present and evaluate the results. According to the evaluation done, the algorithm accuracy level stands around 80% which is a fairly satisfactory rate compared to its simplicity