<![CDATA[The abundance of job vacancy information on various digital platforms often creates inefficiencies in the job search process because applicants must manually select job descriptions, while companies also experience difficulties in assessing applicant suitability quickly and objectively. This study develops a web-based job vacancy recommendation system with a Natural Language Processing (NLP) approach using the Term Frequency–Inverse Document Frequency (TF-IDF) and Word2Vec methods to represent applicant profiles and job descriptions, which are analyzed using cosine similarity. The system was developed using a prototyping method so that it can be iteratively adapted to user needs. Evaluation of the characteristics of the recommendation results was carried out exploratively through analysis of similarity scores and changes in job rankings in several test scenarios. Based on the test scenarios and descriptive analysis of similarity scores and ranking changes, TF-IDF tends to produce lower suitability scores when there are variations in terms, while Word2Vec provides relatively more stable scores due to its ability to represent the closeness of meaning between words. In addition, usability evaluation using the System Usability Scale (SUS) obtained an average score of 89.5 with an Excellent category, indicating that the system is easy to use and comfortable for users. This system is expected to help applicants find relevant vacancies and support a more systematic applicant screening process.]]>
Copyrights © 2026
