Generic framework for recommendation system using collective intelligence

Internet users are in need of a tool that helps them to explore more and more contents on the web. Web users are undergoing a transformation and they are now expressing themselves in the form of sharing their opinions on an item through ratings and reviews or comments; through sharing and tagging content; or by contributing new content. In this changing scenario, recommendation system should not only present contextually relevant items or personalized items but also show items which are hot among other users over the Web. In this paper, we propose an approach that takes users' collective intelligence through their interactions with the contents, their contribution and navigation patterns, and finally suggests best recommendations. The algorithm is independent of the type of item and can be applied to videos, music, photos, news, books, e-shopping products or any other type of items. Proposed recommendation system exploits collective intelligence through user contributed tags, overall community opinion and most common co-occurrence patterns found in users' actions. The performance of the recommendation system has been evaluated through users' tendency of clicking to the recommended items and diversity of the items being consumed by users.