MATURE: Recommender System for MAndatory FeaTURE choices
In the efforts to improve the accuracy of the recommender systems, the usage of the additional information about the user and item rather than just the user-item correlation based on rating becomes an important and integral part. Various algorithms have been developed beyond the mere traditional methods of using user-item correlation, which have exploited the additional user and item information such as the mandatory and discretionary features of the user and various items.
In this research work, we propose a content-based approach MATURE, utilizing the additional information about both the user and item to predict the items for such a user who has current mandatory needs to accommodate as opposed to the past preferences. To the best of our knowledge, MATURE is the first such recommender system that guarantees and ensures the inclusion of the specified mandatory features, when recommending the items to the user and justifies the recommendation by using the information enclosed in mandatory features.
Committee Members: Simanta Mitra (major professor), Johnny S Wong (major professor), Ying Cai, Carl Chang, Adisak Sukul, Udoyara Tim