Social network analysis is an approach to quantify the sociological analysis. It maps the social actors to nodes,socialrelationship between actors to edges,and then uses knowledge of graph theory to solve theproblems in social networks. This thesis puts emphasis upon the algorithm improvement and system implementation in the domain of data mining and social network analysis,on the basis of indepth study on the entity resolution methods based on graph,new similarity measures are proposed and successfully applied to entity resolution on weighted and directed graph,which is similarity measure based on markov random walk. A telcom analysissystem integrated with the proposed algorithms is int roduced and also the proposed algorithms are validated on real telcom data set based on the telcom analysis system.