An approach for extracting tweets from social media factors

News media provides the information to the public about the regular events. Now-a-days social networks, such as twitter provides the user generated data about the news related contents. For this resource to be useful we must cluster the data and provides only the useful information. In this we used density based k-means algorithm and graph clustering algorithm for filtering the data. After filteration we rank the data based on frequency of keywords, relevant key terms and finally based on key term similarity in the dataset. We can also extend to other topics besides news, such as science, technology, sports, and other trends.