Features combination for gender recognition on Twitter users

Gender classification in social platforms and social media has become a relevant topic for the industry because of its impact in making decision process. Gender recognition in Twitter is a business intelligence tool focused on twitter data acquisition, analysis, and process, and it can be used in many ways to transform it into valuable business intelligence data. In this paper, a method for gender recognition in Twitter users is proposed. This method employs several features related to user profile picture, screen name and profile description. This method was evaluated in a dataset with 574 users acquired from Twitter API, these users are located in Aguascalientes City at Mexico and they were manually labelled. The experimental results show an accuracy of 89.5%.