Clustering and ranking social media users based on temporal analysis

Temporal information retrieval is widely used to organize the data according to the time considerations. Social media has rapidly changing time varying content. Temporal information retrieval has made it easy to explore this kind of dynamically changing content. Clustering involves grouping the data in such a way that their inter cluster similarity is low and intra cluster similarity is high. Similar set of data belongs to a single cluster making analysis of data easier. In this work, relevant features are extracted from the Social Media platform `Facebook'. These features form the basis for ranking and clustering of users. Proposed ranking technique involves the temporal dimension of data and comments by different users. Data mining toolkit WEKA is used for clustering. Three algorithms used for clustering are: XMeans, Expectation Maximization and Hierarchical. Comparison of results for these algorithms shows that XMeans and Expectation Maximization are better for Social Media based Temporal data rather than hierarchical clustering algorithm.