Folksonomy Based Fuzzy Filtering Recommender System

As the choice of items on the World Wide Web increases, the difficulty to choose the right item also increases. Fast and effective recommender systems help to overcome this difficulty by automatically using algorithms and data to recommend items like movies, webpages, videos, etc. Conventional collaborative filtering recommender systems use the rating data to find the neighborhood of users in terms of their preferences. However, these recommender systems do not incorporate the domain knowledge in their recommendations. In this paper we propose a hybrid folksonomy-based fuzzy collaborative filtering recommender system which uses domain knowledge of the items and combines them with the collaborative process. To further enhance the performance of the system we use fuzzy clusters to represent the neighborhood in a probabilistic manner which makes the recommendation process more accurate as verified by our experimental results