Gender Prediction from Social Media Comments with Artificial Intelligence

Özer Çelik, Ahmet Faruk Aslan
2019 Sakarya University Journal of Science  
In the 21st century, which can be termed as age of artificial intelligence, machine learning (ML) techniques that can become widespread and improve themselves can be given more quality services to humanity in many fields. As a result of these ML developments, nowadays many companies use predictive models to estimate customer behavior. Also, with increasing use of social media, the companies have started to deliver their products and services to their customers via social media accounts. But
more » ... y customer is not interested in all product or service. Each customer's area of interest is different. Gender is one of the main reasons for this difference. If the gender of a social media user is determined correctly, the amount of sales may be increased by offering the appropriate products or services. The main aim of our study is an estimation of genders of the commenters thanks to machine learning techniques by analyzing the comments of companies posting on Facebook. In context of the study, the genders of the commenters labelled based on commenters' name. The data set is divided into training and test data as 70-30%. As a result of the study, it was seen that machine learning methods predicted with similar accuracy rates, while the highest accuracy rate (74.13%) was obtained by logistic regression method.
doi:10.16984/saufenbilder.559452 fatcat:ipn77plj5bddrmiimef7sysh4i