Recommending Domain Specific Keywords for Twitter

Twitter has become the most popular social media in today’s world. More than 284 million users are online monthly, and 80% user accesses their twitter account through mobile. A tweet is limited to 140 characters, so it contains concise information about particulars. Due to its popularity and usage, near about 500 million tweets are sent per day that relates to different domains. This work focuses on recommending domain specific keywords for twitter. For this purpose, 10 domains are chosen as a sample. Then we apply Term Frequency-Inverse Document Frequency (TF-IDF) and Log likelihood methods and compared the keywords extracted from both against each domain to make our result much valuable. Furthermore, the categorization of keywords is made as noun and verb, and also finds out the sentiment words. At the end, a relevancy test is performed from five users. These keywords can be great value in clustering tweets data and can be used for identifying a user’s interest in any specific domain. Furthermore, these keywords are of the great asset for advertisement purpose.