Machine Learning for Older Jobseeker and Employment Matching

Abstract

The challenge of the growth of the aging society is causing a critical economic burden that is affecting the gross domestic product of many countries. Normally, recruiters for information technology employment face the problem of matching appropriate candidates with available positions as they may have insufficient knowledge to determine whether candidate profiles are suitable. The requirements and demand for jobs may not be enough to advance the situation while matching skills and competence with job descriptions often preclude older candidates. This research is supported by the Thai Gerontology Research and Development Institute (TGRI). Herein, we present a solution involving data mart modeling and machine learning. In the data mart modeling, an extract, transform, and load (ETL) process that pulls data into a data mart (star schema) and then applies distance measures to calculate matching. Historical records of both the companies and older job seekers are kept in the data mart to understand behavior and keep track of data. Moreover, we collect disease information and map to ICD-10 code. This can help employers to find appropriate job seekers related to their skill sets and limitation of disease. We treat their skills as terms and employment positions as documents and then calculate the ranking of each job profile. The results show that the outcomes garnered from the distance measures are consistent and creditable.