Recommendations through click stream: Tracking the need, current work and future directions

Abstract

Recommender system is a tool that provides suggestions to customers. Recommendations are provided for the products that a customer may like in future or that are close to the target customer. On an e-commerce website good recommendation plays an important role for the seller and the buyer. So far researchers have digged out many methodologies for recommendation that may use explicit ratings or implicit data. Keeping track of customers surfing behavior can also help in endorsing products to similar users. Finding preference levels of a product for a particular customer can provide accuracy in recommendation. In this survey, we review recent developments in recommender systems based on click stream data and discuss the major challenges faced. We compare and evaluate available algorithms and examine their roles in future developments. We will discuss the methodologies and techniques that the researchers have devised for e-commerce websites with their drawbacks and a relative comparison of their performance.