Recommender Oriented to Complicate Task in E-commerce

Abstract

Recommender systems are used by e-commerce sites to suggest products to their customers and to provide consumers with information to help them determine which products to purchase. Traditional recommender is oriented to a unique recommendation algorithm based on clear preference. Unfortunately, it is not enough. In a lot situations, user hope recommender help them to finish a complicate task based on fuzzy preference, not only to provide candidate recommendation list. In the paper, an advanced recommender, recommender oriented to complicate task (ROCT), is proposed to meet the requirement. A general design of ROCT is presented, and its architecture and working procedure is detailed. In ROCT, soft constraint technique is used to model userpsilas preference and critique example is used to elicit userpsilas preference.