An Approach for Predicting Links in Social Network using Modified C4.5

Now days in the modern society online social network plays an vital role, link prediction is one of the most vital function on the online social network. There are variety of software of prediction of links which includes prescribing new things is available in market to users, friendship recommendation and finding spurious institutions. Prediction of link is a critical issue that provides to network evolution. The assignment is to evaluate the probability of the links of connections in future in a particular system. Social community relates humans through some kinds of connections. Therefore, it is all the more intriguing to predict the presence and the kind of a future link. This check the predicting the links in social networks because these latter allow concurrent institutions with a few types. On this gadget latent characteristic representation technique is used for representing the features from the consumer profile and KNN classification is used for link prediction. We archive parallellism using multithreading with parallel C 4.5 algorithm. This device will predict a friendship link on a social community on the premise of description or biography given inside the profile by means of the user