Abstract:
The Sri Lankan foreign employment sector faces challenges in matching unskilled, low
skilled, and semi-skilled job seekers with appropriate foreign employment opportunities. The
current recruitment process is inefficient, lacks transparency, and often fails to meet job
seekers’ needs, leading to high turnover rates and dissatisfaction. This research aims to
address these issues by developing a job-matching platform that leverages machine learning to improve the accuracy and relevance of job recommendations for Sri Lankan workers seeking foreign employment.
To improve job-matching accuracy, the project uses a hybrid machine learning strategy that
combines collaborative and content-based filtering algorithms. The platform can process job seeker profiles and job openings using this methodology, producing recommendations that are specifically suited to each user. The design and development stages were guided by
requirements gathered through surveys, interviews, and document analysis, guaranteeing that the platform is usable, accessible, and culturally appropriate for the intended audience.
Preliminary results from testing demonstrate the platform’s ability to provide accurate job
matches, with key metrics such as the Confusion Matrix, F1 Score, and AUC-ROC used to
evaluate its effectiveness. Initial testing shows promise in that the platform can increase job
placement rates and meet user needs. This platform provides a scalable solution that fits well with the changing Sri Lankan job market and foreign employment demands.
