Location Inference for Non-geo tagged Tweets in User Timelines

Tags: , , ,

Aim:

To obtain the information of user location from the tweet, which are not with the help of geo-tagged, but with the help of machine learning.

Existing System:

Business can obtain the geo location of their user with the geo-tagged content like images or tweets. But not all tweets were geo-tagged. The existing system depends on the provision of the content to get the user location. The user provided content is not always contains geo-location.

Proposed system:

To overcome the fallback in the existing system we propose a machine learning based programmatically identify the user tweet country. Our system uses the users text content and other properties of the tweet like time zone, language etc. to identify the tweet country. An SVM model is trained with given dataset. This model is used to predict the country of origin on the tweet.