An Efficient Link Prediction Model Using Supervised Machine Learning

Link prediction problem is subsequently an instance of online social network analysis. Easy access and reach of Internet has scaled social networks exponentially. In this paper, we have focused on understanding link prediction between nodes across the networks. We have explored certain features used in prediction of link using machine learning. The features are quantified exploiting the structural properties of the online social networks represented through a graph or sociograph. Supervised machine learning approach has been used for classification of potential node pairs for possible links. The proposed model is trained and tested using standard available online social network datasets and evaluated on state-of-the-art performance parameters.