Intern retrieval from community question answering websites: A new variation of expert finding problem

Abstract

In recent years, with the advent of new scientific and technical areas, hiring suitable interns has become very important for large companies. In internship programs, to reduce the risk of unsuccessful recruitment, companies are actively seeking those candidates who are ready to become skillful experts, and also incur the least financial costs. Therefore, generalists (i.e. Hyphen-shaped people) are the most suitable candidates for such positions. These candidates have general knowledge in the required skills of the position and do not have expertise in any other area. One of the environments that accurately reflect the knowledge of people is community question answering (CQA). This study is the first that focuses on retrieving interns from CQA websites. It uses the concepts of generalist and shape of expertise to identify suitable candidates for the internship programs of companies. Specifically, in this paper, we define the intern retrieval problem, in which given a set of required skills of an internship position, a ranking of candidates is generated so that the generalists who have general knowledge in those skills are retrieved in top ranks. We propose two retrieval models to address the problem. In order to evaluate the performance of these models, we introduce two specific measures (i.e. coverage and optimality). Our experiments on three test collections extracted from StackOverflow demonstrate the effectiveness of our models in comparison with several baselines.