Provide a Multi-Factor System for Social Media Mining Relation

Abstract

The high volume of data on social media leads to some features such as high quality, big data and direct reflection from real human society. It is important to anticipate a missing or unrelated link on current social networks and a newly added or deleted link on future social networks. Calculating the similarity between a pair of nodes is a good way to explore communications. In a social network, a node usually contains features such as the characteristics of Internet social networks. This information can be used directly to calculate the similarity between two nodes. There are many metric based on network topology. These metric are defined based on graph theory. The data used in this study is called Pokak, which is one of the databases published at Stanford University. The proposed method in this paper is to explore the relationship between users using a combination of node-based metric and network-based topology metric. In the calculation section, the similarity between users according to the metric based on node, we use the existing data clustering. The k-means method is used for clustering. Users who have similar features are placed in a cluster. In the second step, we use topology-based metric. If the similarity is greater than the threshold, the proposed algorithm for these two nodes is considered a friendship line.