In cyberspace, Social networks have been born as a new type of websites and have gained an enormous range of users and fans. Social networks are one type of social media and are places for forming virtual communities of interested users. Internet users have been classified in different ways based on their type of using social networks. This study seeks to provide a mechanism to predict patterns of behavior in social networks. Due to the expansion of social networks, the selected network requires a model based on the new strategic decisions or policies for better serve users. This study uses data mining techniques for classification and analysis of social network users for better understanding of their behavior and improving services and developing appropriate strategies. Understanding behavioral patterns of users of social networks lead to better adaptation to user needs. The user population applied for analysis includes 31033 users that use a specific Iranian Social Network regularly. A method for clustering and orientation analysis based on past users behavior using CRISP-DM and data mining software is deeply analyzed and described. A full perception of users’ behavior will result in a better match of social network features with users’ needs as well as a high value added for users and profitability for social network owners.

Keywords: Social Network, Data Mining, Clustering, Davies Bouldin, Users Behavior

Full-Text [PDF 931 kb]   (8744 Downloads)