Community detection using fractional weighted K-core graph for facebook data

A social website that focuses the formation and reflection of social relationships between a set of people with shared interests. Users can also exchange ideas, activities, events, and hobbies within a community. Because of the very dynamic nature of social media data, the important aspect of community detection merits additional consideration. The establishment of social media networks results from the affiliation of entities through their social interaction. The finding of social media communities is made possible by analyzing the structure of such networks. Identifying the most influential users and the most strongly connected neighbors inside communities is a critical issue in the social network. The proposed approach is implemented on seven datasets of pages retrieved from the Facebook search API, and the results are found to be the best among the available methods on the specified datasets. It demonstrates that the coreness of the users in the discovered communities is measured by user interaction activities, which yields successful results in identifying the highly influential individuals. Accuracy and Z-score are used to determine the best performance of the community coreness. The Fractional Weighted K-Core Graph (FWCG) score is used to measure community coreness and detect influential communities. Three evaluative indexes of accuracy, recall and F1 score are considered to estimate overall influence of the proposed work and given significant results.