Link Prediction in Weight Social Networks Based on Time Series

Link prediction is an important task in social network analysis, the core of which is to predict the establishment of future links between nodes in social network. A weighted network model based on time information is proposed. Because the preference of nodes in the network will change with external factors, the time decay factor is introduced into the algorithm through the logistic function. The link weights between node pairs are determined by the link establishment time information, and the weighted network is used to improve the effect of supervised learning link prediction. Experiments were carried out on the academic co-author network, and the weighted network model was constructed to predict the links using supervised learning technology. The results show that this method is superior to the traditional weighted network method in Accuracy, f-measure and AUC.