Open AccessCommunity Detection under Stochastic Block Model Likelihood Optimization via Tabu Search –Fuzzy C-Mean Method for Social Network Data
Author(s) -
Ali Falah Yaqoob,
Basad Al-Sarray
Publication year - 2020
Publication title -
iraqi journal of science
Language(s) - English
Resource type - Journals
eISSN - 2312-1637
pISSN - 0067-2904
DOI - 10.24996/ijs.2020.61.10.27
Subject(s) - stochastic block model , tabu search , bayesian information criterion , akaike information criterion , cluster analysis , data mining , mathematical optimization , computer science , mathematics , community structure , minimum description length , machine learning , artificial intelligence , statistics
Structure of network, which is known as community detection in networks, has received a great attention in diverse topics, including social sciences, biological studies, politics, etc. There are a large number of studies and practical approaches that were designed to solve the problem of finding the structure of the network. The definition of complex network model based on clustering is a non-deterministic polynomial-time hardness (NP-hard) problem. There are no ideal techniques to define the clustering. Here, we present a statistical approach based on using the likelihood function of a Stochastic Block Model (SBM). The objective is to define the general model and select the best model with high quality. Therefore, integrating the Tabu Search method with Fuzzy c-Mean (FCM) is implemented in different settings. The experiments are designed to find the best structure for different types of networks by maximizing the objective functions. SBM selections are computed by applying two types of criteria, namely Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC). The results show the ability of the proposed method to find the best community of the given networks.