GenderPredictor: A Method to Predict Gender of Customers from E-commerce Website
Siyu Lu, Meng Ya Zhao, Hui Zhang, Chen Zhang, Wei Wang, Hao Wang · 2015
While e-commerce has grown substantially over last several years, more and more people are utilizing this popular channel to purchase products and services. Thus the ability to predict user demographics, including gender, age and location has important applications in advertising, personalization, and recommendation. In this paper, we aim to automatically predict the users' genders based on their product viewing logs. Our study is based on a dataset from PAKDD'15 data mining competition. We propose an architecture for gender prediction, which consists of the "machine learning model" and the "label updating function". The experimental results show that our proposed method significantly outperform baseline methods. A detailed analysis of features provides an entertaining insight into behavior variation on female and male users.