link prediction in social networks using improving adamic-adar and jaccard methods

in recent decade, social networks are an important issue and those are a part of people life. the number of connected users are increasing daily. active users in social networks are more than one hundred million. people in social networks interact together and share their life. some example of social network are facebook, google+, instagram and etc. link prediction is an important and challenging problem in social networks. social networks are dynamic and it give the importance of link prediction. dynamic means that it is maybe members and relations not exist in next moment. on the other hand maybe in the next time will create new member or new relation. so there is a need to predict not exists links at this moment that maybe will create in the next moment. we need some information to link prediction, that we can gain those from the network graph. forasmuch as adamic-adar have good precision on the different datasets, in this paper, we improving adamic-adar method using jaccard index. to evaluate proposed method, we compared it with several link prediction method. the proposed method was successful that increasing the adamic-adar precision of facebook's dataset 0.2%, and the rate of improvement on the hamster dataset is 23%.