Twitter data analysis by means of Strong FlippingGeneralized Itemsets

Abstract—Social Network Services (SNS) like Facebook, Twitter, Sina Weibo etc. are widely used all over the world today. People use them to keep in touch with friends and make new friends online. In China, Sina Weibo is one of the most successful social network service and has great influence to peoples’ daily life. Until the end of 2015, Sina Weibo has more than 236 million monthly active users and the revenue reached 478 million US dollars. Targeted marketing for a very important aspect for SNS operators and tired party advertising companies. In this research, < Final Year Projects > k-means based clustering has been carried out for grouping Weibo users according to their features. The grouping results can be used by Sina or third party companies for target marketing with differentiated customers. The official API provided by Weibo open platform was used for collecting data. Traditional K-means and an improved K-means algorithm was used for clustering. The classified user groups show strong independent characteristics. The classifying method and results are valuable for further SNS analyses and targeted marketing.