A social network is a social structure made up of a set nodes, which represents social actors (such as people, organizations), and edges or lines represents relationship between these nodes or actors. Social networks have important roles in the dispersal of information and innovation, the analysis of such networks, attracted much attention in the research area. The analysis of social network can be done as a whole, which means the representations of all of its actors and identification of structures, present in that social network, that lead to the presence of communities. In the method of community detection, the main aim is to partition the network into dense regions of the graph, and those dense regions typically correspond to entities which are closely related, and can hence be said to belong to a community. In any complex network, communities are able to exchange and offer information because members in one community have similar tastes and desires. The determination of such communities is useful in the context of a variety of applications in social-network analysis, including customer segmentation, recommendations, link inference, and vertex labeling and influence analysis. This paper presents a survey on community detection approaches, which have already been proposed, and also discussing the type of social networks on which those proposed approaches are applicable. This survey can play a significant role in the analysis and evaluation of community detection approaches in different application domains.

No abstract

No abstract

Social networking websites have become an easiest way to make the common people thoughts and reviews to become public. Among those websites, Twitter data's are in boom, because of heavy interests of people to update their information in that website. Detection of communities for Twitter data has already been done by the other authors, but still communities detected with high strength or quality are lagging behind. In this paper, the data collected from Twitter have gone through sentiment analysis and the final scores of that analysis have been used for the plotting of the graph which acts as an input to the community detection algorithm. The twitter data's communities were detected with the detection of noise too, and upon removal of those noisy data, the strength of the detected communities used to get increase. The detection of the outliers or noise has been done with the help of DBSCAN algorithm and the communities have been detected by Newman Girvan algorithm. In this study the proposed sentiment analysis algorithm and the community detection technique have been successfully implemented and evaluated. The results from the collected data sets from Twitter have shown the communities, which were properly detected with the help of the proposed methodology. The communities were actually grouped according the sentiment scores derived and the number of words, for each tweets. Each community shows the connection according the high and low sentiment scores. The quality of the detected communities has been measured by centrality, modularity and conductance and has been compared with four other community detection algorithms i.e. with Louvain, Walktrap, Leading Eigenvector and Fast Greedy algorithms. The results were positive in maximum times when compared on the basis of the considered metrics with the other community detection algorithms.

No abstract