Application of Community Detection Technique in Text Mining

The word community detection indicates the group of similar kinds of objects. In data mining techniques, this term is used as the cluster formation or data cluster analysis. The presented work is focused on evaluation of text for finding the similarity among them in order to find the different context of data. According to the definition, community detection is an approach of data clustering and their graphical representation. Therefore, in this paper a community detection technique is presented for mining text data. Due to lack of different nature of data influence, in this work the social media text data is considered for experimentation. The proposed data model first analyzes the text document, and then the features of data are computed. These features are used further for finding their similarity among two instances of data. Based on their similarity, the data is clustered and their visualization in terms of community groups is performed. The implementation of the proposed technique is performed using JAVA technology, and their performance in terms of precision, recall, and F-measures is performed. The performance of the proposed technique demonstrates that the technique is able to perform clusters efficiently and is also able to find the partially similar data instances between two clusters. That enables it to provide more accurate cluster formation such as fuzzy C-means clustering.