Research Information System
Thesis Type: Postgraduate
Institution Of The Thesis: Marmara University, Institute of Social Sciences, Department of Business Administration (Eng), Turkey
Approval Date: 2024
Thesis Language: English
Student: OSMAN KAVAFOĞLU
Supervisor: Hüseyin Ekizler
Abstract:
Development of the technology and resources have increased the importance and value of human resources management for organizations. Human resources management encompasses functions such as recruitment and selection of staff, wage administration, performance management, benefits governance, professional growth and education, strategic human resources management, job market analysis, and staff retention and separation. The study focuses on the recruitment function of human resources management. Human resources technologies have become a fast-growing industry, applying artificial intelligence and machine learning solutions to improve and facilitate its processes. Matching algorithms are artificial intelligence solutions that aim to match job listings with candidates in order to facilitate the recruitment function. The aim of this study is to investigate the importance of algorithms that can be used to match listings with job seekers on a recruitment platform, based on given criterion from the perspective of subject-matter experts. To this end, on a recruitment platform, managers of marketing, sales, product, customer experience and software development departments who are experts in the field have been scaled to identify the importance of alternative algorithms and criterion. The data obtained were analyzed using the AHP, SWARA, VIKOR and WASPAS methods of multi-critical decision-making and statistically tested with Kendall’s Coefficient of Concordance.
In the first stage of the study, the importance levels of the criteria according to subject-matter experts were determined by the AHP and SWARA methods. In the second step, with comparison, preferences of subject-matter experts for matching algorithms regarding the criteria are ranked utilizing VIKOR and WASPAS methods. In order to determine whether utilized methods caused any difference in ranking results, Kendall’s Coefficient of Concordance test is applied. As a result of the analysis, according to the experts participating in the study, the most important criterion was the User Satisfaction, while the best algorithm regarding the criteria in the study was Natural Language Based matching algorithm. Kendall’s Coefficient of Concordance showed that despite different MCDM methods, with this data set, results would be the same.