Machine Learning Based Product Classification for eCommerce

Abstract

ABSTRACT Product research in web shops involves categories as a helper system for search and browsing. The categorized offer reduces user’s mistakes and misled presentation. Hence, learning a product classifier to work with high precision and proper recall is of fundamental importance in order to provide high-quality shopping experience. However, products can be perfectly described they can miss users’ vocabulary when they do research. Thus, users’ content in form of search phrases has been utilized in order to extent a product data. Thus, original product names and classes have been supplemented with users’ search phrases, which changed classification accuracy in significantly positive direction but also slightly influencing on data consistency. The article presents new knowledge, research framework, machine learning classifiers selection and result analysis with implication for academia and some suggestions for business practitioners.