Profiling Social Media Users, a Content-Based Data Mining Technique for Twitter Users

Trends in Social Networking have brought millions under a single roof. The mode of information sharing by people around the globe has been occupied by chat, blogs, tweets and so on, using different social media sites. To profile the user with respect to the topic(s) discussed in social media, "Profiling Social Media Users (PSMU)" algorithm is proposed in this work. Using NER approach, the keywords in the posts are clustered into seven predefined clusters and further, unique clusters are formed based on synonymous of keywords in posts using WordNet approach. Users are profiled based on having highest number of keywords in respective clusters. Experiments conducted on real world data set from Twitter Social Media show unique clusters being formed by PSMU and profiles the Twitter Users accurately by achieving 97.53 percent accuracy.