Header menu link for other important links
Mining social networks to discover ego sub-networks
A. Madani,
Published in Institute of Electrical and Electronics Engineers Inc.
Pages: 135 - 139
A decade ago, no one would imagine that social networks would witness such rocket development to take part of our daily life. Within relatively short period of time, social networks gained global attention by a huge number of users. Since then, social networks grow bigger to accommodate millions of users. One challenge though is to automate the process of categorize friends into social. Therefore, several data mining solutions were introduced to automatically attempt to classify ego networks, however, issues related to categorization accuracy as well as the efficiency and effectiveness to classify ego networks are still lacking. In this paper, we propose a method that mainly utilizes features extracted from network topology and node profile to better identify social networks. We believe that extracting comprehensive and meaningful features to be used in conjunction with a roust classifier would increase the accuracy level of ego sub-networks detection. The proposed method was tested on a benchmarked data set and the performance was compared to two state-of-the-art methods namely Maximization Likelihood Like and Enhanced Link Clustering and showed valid improvement. © 2016 IEEE.