Essay --

Keywords—recommender system; fuzzy system; social matchmaking; crisp set; fuzzy set I. INTRODUCTION In the ‘information era’, one of the key problems is to deal with more information than to practice to make practical decisions. User is bombarded with information whether or not he positively looks for it. Recommender systems are designed to help individuals to deal with this information overload problem and enable them to make evaluative decisions [1]. Traditional RS provides items, information and services to the user. These items are like products, movies, CDs, music, news, books etc. Tapestry [17] is the first manual RS and Usenet newsgroup launched by GroupLens is the first automatic collaborative filtering RS [6]. The most popular existing recommender systems are Amazon.com for e-shopping [7], MovieLens recommending movies, news by Googlenews, music at Pandora, EntreeC giving restaurants [11], CDs at CDNow [18] etc. In many past years, for building recommender systems various approaches have been developed that utilize non-personalized, demographic, content based , collaborative filtering, knowledge based and hybrid [11]. Evolved research areas like social matchmaking RS enable people to people matchmaking [2] like matrimony system recommends bride to groom and vice-versa. Using such systems, users can meet the other individuals of complementary needs like getting jobs (employee-employer), college admissions, mentor-mentees, student helper, addressing community issues, solve technical problems and counseling [3]. In social matchmaking systems, successful reciprocal recommendation occurs where two users find each other based on their complementary needs. For example, a bride finds the ideal groom, and the same groom li... ...= â€Å"Very Low† (0.2) The sample of recommendations for the active lady is shown in TABLE IV. The snapshot of the result for same expectations is given in Fig. 2. The system is not providing the partners who having â€Å"Low† value for crisp sets (religion, caste, occupation, diet, smoke, and drink). The experiments are observed for ten users and precision, recall, F1-measure is calculated. For getting these values, recommended results are used. The average of precision, recall and F-score are 79.45%, 85.65%, 82.43% respectively. V. CONCLUSIONS This paper focuses on Partial Fuzzy Recommender System used for matrimony in the context of the Indian society. This system addresses the abundance of information and directs users to precise data requirements in terms of matches, eliminating irrelevant information. Recommendations can be further improved for reciprocity.