Kimani, Daniel Kamau

Kimani, Daniel Kamau

An enterprise integration and service developer with over 5 years working in the corporate world.

Project Summary

Project Title: A Market Basket Analysis Model to Address Visitor Cold Start Prediction Using Association Rules

Research Supervisor: Prof Elisha Opiyo

Abstract: E-commerce websites use recommender systems that recommend products with the highest user rating and those products that have similarities during customer interactions. These techniques give results from information retrieved during customer search and can be interpreted to match what the use is looking for. The recommenders simply respond to queries that customers invoke on the websites rather than customizing their responses to customer’s need. These recommender systems face multiple challenges like the cold start problem where a new user tries to purchase from the system or there is a new product that has no rating at all and therefore lack of enough information to recommend to the customer or the product, and an issue with overspecialisation. MBA approach with AR was proposed in this research to address the problem where the user has not interacted with the system before or there is insufficient information on the customer profile. The scope of the presented approach was to come up with a model that could accurately produce useful recommendations to the customer by identifying relationship patterns between products. The approach used Association Rules and predictive analysis of the products and intelligently identified similar sets of products that may interest a user so as to suggest those products that would satisfy a consumer more.