DESIGN AND DEVELOPMENT OF SOCIAL COMMUNITY FRAMEWORK USING DATA MINING

Abstract

With the dawn of the living web 5.0 and Ontological semantic networks, open source social interaction Platform popularity and dependence has gained attention of researchers for both industry and academician. This new area of research focuses on social behavior of netizens. A digital avatar along with net information about user’s choices like membership of sports group, financial group, political or entertainment society, video gaming society etc. is available processing and recommending different products, services and friends. The main challenge associated with a Recommendation System is to recommend useful information to the user at right time. Friend Recommendation, which is the one of the indispensable feature of Social media, has taken it to new height. Facebook, Twitter, LinkedIn, MySpace have captivated millions of users now a days. But the antecedent research work on Friend Recommendation cynosure on user current relation in Social Networking. Facebook, one of the most prominent social networking platforms provides the personalized friend recommendation based on FOAF (Friend of a Friend) ontology. MySpace is based on PYMK (People You May Know) friend recommendation. Basic perception behind it is that probability of a person knowing a friend of friend is more than unknown person. This work proffers a unique approach of friend recommendation based on the user’s interest and based on user current location. The proposed system recommends friends based on user interest. Further, user interest keeps on changing. To overcome this challenge, recommendation System is proposed using Ontology and Spreading Activation. we developed a recommendation system which provide content recommendation to user based on user interests which gets ix changes over the period of time and system learns this using the spreading Activation algorithm. User interest is being captured using the Spreading Activation. Spreading Activation has been used to overcome variation in user interest. Our experimental results have shown the benefits of considering Spreading activation and ontology in friend recommendation in as social networking.