Multi-Criteria Based Recommender System Scalability Optimization : The Approach Based On Clustering of Users

The era of Internet & the ever blooming E-Commerce applications have exposed people to the mines of information and variety of products to choose from while shopping on-line. In such situation it is quite obvious to get lost in exploring abundance of products and related information. This causes the waste of large amount of precious time and may at the same time, create confusion in the minds of prospective on-line shoppers regarding purchasing decision to be made, leading to in-vein search and unfruitful buying attempt. This is categorized as the information overload problem. Recommender systems play an important role in such scenario by presenting the buyer with the ranked list of products preferred by existing buyers for the same/similar products. The number of features or attributes of the product used to determine the order of the product in the ranked list will define the approach used by Recommender system. Traditional approach is called as single criterion based approach for recommender system while the other one is multi-criteria based approach for recommender system. As the number of users of the e-commerce site increases, the time required to form ranked list will also increase with the great amount in case of multi-criteria based recommender systems. This requires optimization criteria to be applied to minimize the time required to form ranked list while maintaining the accuracy of the result.