Date of Publication :17th May 2017

Abstract: The Demographic property of clients such as gender, age, and education gives important information for e-commerce service providers or specialists in merchandising and personalization of web applications. However, online clients often do not give this type of information due to privacy and security related reasons. In this we proposed a method for previsioning the gender of clients based o0n their catalogue viewing data on e-commerce systems, such as the date and time of access, list of categories and products viewed, etc. We use a machine learning techniques and investigate a number of characteristics derived from catalogue viewing information to prevision the gender of viewers. Experiments were carried out on the datasets. The results 81.2% on balanced accuracy shows that basic characteristics such as viewing time, products/categories characteristics used together with more advanced characteristics such as products/categories sequence and transfer characteristics effectively facilitate gender prevision of clients

Reference :