Multi Criteria Decision Making (MCDM) Study for Job Market Place Recommender System

• Main Author: AKHIRO (NIM 23506038), RIDHO
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/9079
• Institution: Institut Teknologi Bandung

<br /> <br /> <br /> Indonesia is a country with many labor force. The number of labor force urges the existance of job market place which is a place where job applicants and job providers (companies) who offer job vacancies meet. The problem that emerges is that sometimes it is difficult for applicants to decide which job vacancies to apply and the job providers themselves sometimes having difficulties to find the right candidates. <br /> <br /> <br /> <br /> This thesis attempts to employ recommender system to solve such problem. This system is expected to help applicants or job providers to receive good recommendations. This thesis uses content-based recommender system and multi criteria decision making (MCDM) technique for providing the recommendations. <br /> <br /> <br /> <br /> The thesis process is started with literature study about recommender system, decision making theory, and theory about employee selection. Furthermore, problem analysis in order to apply recommender system in job market place is conducted. MCDM method selection is conducted and the Weighted Product Model (WPM) is choosen as the selected method. <br /> <br /> <br /> <br /> Evaluation of the designed algorithms is conducted by experiment which use data consist of curriculum vitae (CV) and job vacancies advertisements. The results of the experiments show that the WPM-based algorithms can be used to generate recommendations better than the Weighted Sum Model (WSM). Nevertheless, the experiments also show a special case in which the system may produce bad recommendations. <br /> <br />