Location estimation of non-geo-tagged tweets.

Internet users are getting more and more dependent for information regarding their daily lives. Most of the users are connected to each other using social networks. Social networking sites not only helps the users to connect and talk to each other but also share information with each other. Twitter [1] users attach their location information with the post or tweet to show their presence at the location. But, not all users tags or integrate the location information within the post. If a person wants to obtain the latest updates about an event then he/she have to go through all the tweets about that event, which is impossible because nearly 500 million tweets are posted on Twitter on a daily basis. Using Twitter the users can post up to 140 characters in their posts or tweet. Also, the tweets that originate from the location of the event are latest and contain new facts and the rest of the tweets convey that information only. Non-geo-tagged tweets are eliminated by the traditional systems. This paper presents a method to tag the non-geo-tagged tweets with the location then the user would be able to obtain the latest information by including the new. The proposed method performs better than previous methods and yields better results. [ABSTRACT FROM AUTHOR]