| dc.description.abstract | The study proposes a novel suggestion of a recruitment support system based on
artificial intelligence (AI) to augment the security, equity, and efficacy of international
job recruitment of foreign employees. The migrant workers are likely to have dubious
job information, corrupt recruiting firms, lack of expertise in the activities, and language
barriers that increase the risk of exploitation and poor working outcomes. The proposed
system is based on the four main components, which are: employment congruence,
prediction inability of the skills, detecting the fraudulent agency, and multilingual
chat-support as the tool of increasing the transparency and informed selection. It
was tested on 2000 profile of workers, 1500 overseas jobs and 300 recruitment agents
with 80/20 split, which is 2000 workers profiles, 1,500 foreign job openings and 300
recruitment agents. The accuracy of the job matching prediction by the use of the
TF-IDF-cosine similarity (Top-1) was 0.86 and the mean absolute error of the skill
mismatch prediction using the regression analysis was 0.41. It was trained on machine
learning classifiers, which classified fraudulent recruitment agencies with F1 score of
0.89 being discovered utilizing confirmed agency records and complaints. Regardless
of the limitations linked to the quality of data, incomplete worker profiles, and language
barriers, the results suggest that AI-based hiring systems have great possibilities of
reducing the possibility of exploitation, better access to reliable information, and safer
and fair overseas hiring. | en_US |
