Exploring Graph-Based Techniques in Job Recommendation Systems

The transformation in working styles in recent years due to a rapidly changing world has popularized job application platforms, which has in turn led to a vast amount of data that can be used to guide job seekers to job advertisements. This paper explores different methods for modeling the interactions between users and items, which in this study corresponds to job seekers and job advertisements respectively; hence, we develop an item-based collaborative filtering recommendation system through alternative graph-based methods. We evaluate the proposed methods on an interaction snapshot of the dataset of a Turkish online job application platform, as well as with the preference confirmation of real users who work on the same platform as HR experts.