Analyzing Evolving Trends on Social Networks by using Temporal Bipartite Networks

People's online social media activity when requesting a post is a random variable, resulting in a highly skewed ordering of postings. This research work discusses about the reality of people's interest in a post, which might increase or decrease exponentially or linearly. Based on people's evolutionary interests, this study proposes a Growth-based Popularity Predictor (GPP) approach for estimating and ranking web-contents. Three different types of web-based real datasets are used to assess the performance of proposed model, they are: Movie lens, Facebook-wall-post, and Digg. The performance fo the proposed systems is also assessed by using four data gathering measures. The received operating qualities include locality novelty, AUC, Kendal's Tau, and precision. According to the findings, performance forecast can be enhanced much more if the score is transformed into a cumulative projected item's rating.