Grouping of users for social network (SN) analysis is an important process that assists in understanding the behavioral patterns. Numerous machine learning clustering methods are exploited to group users in SN. But, the general purpose similarity algorithms grouping is depending on behavior encounter variety of issues owing to intrinsic nature of users‟ data in SN. One foremost reason is constraints have to be considered while grouping users in SN. Other reason is prerequisite of capturing huge amount of information about users that involves superior amount of computational complexity (CC). This paper proposes a parallel algorithm based on Map Reduce by using Hadoop for identifying group of social network, called, Map Reduce-based Weighted Multi-Constraint Correlated Clustering (MR-WMCC) method. Here, users communicate with each other with the objective of examining behavior by considering the actual tweet list of the user based on hash tag. This is called as Map Reduce-based Tag Identification (MRTI) model. MRTI model speeds up the computational process through parallelism, therefore minimizing the dimensionality. Thus, the CC involved is evaluating the similarity among users depends on behavior is lessened. With dimensionality lessened tweets, an effective Weighted Multi-Constraint Correlated Clustering (WMCC) model depends on correlation measure with consideration constraint among object and cluster and their significance in SN is presented. Depends on the occurrence of constraints among object and cluster, every constraint‟s significance is estimated.
Volume 11 | 06-Special Issue
Pages: 1366-1376