Scholarlink

The selection process after the recruitment stage becomes a challenge for recruiters in verifying the suitability of candidates’ qualifications for the required positions. Currently, the selection process still relies on Microsoft Excel as a tool for keyword searches in candidate data, requiring approximately 30 s to process a single keyword. The implementation of the Content-Based Filtering method in the applicant recommendation system is highly suitable for enhancing the recruitment process efficiency, as it focuses on analyzing the similarity of descriptions in the curriculum vitae data uploaded by candidates. This system operates by calculating the similarity of candidate data based on the entered keywords. The system development process includes several stages, such as data preprocessing, Term Frequency–Inverse Document Frequency (TF-IDF) calculation, Cosine Similarity computation, and ranking the results based on Cosine Similarity values from highest to lowest. The experiment results show that searching for a single keyword takes approximately 0.7 s, which is 29.3 s faster than the current candidate selection process. Additionally, a user satisfaction survey evaluated using the DeLone and McLean model, achieved a score of 95.4%, indicating that the system effectively enhances the accuracy and efficiency of applicant recommendations in the recruitment process.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Boosting Job Matching Accuracy: Implementing Content-Based Filtering in Job Applicant Recommendation Systems

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Rasyid Alim Aulia,
  • S. Y. Yuliani,
  • Angga Aditya Permana

摘要

The selection process after the recruitment stage becomes a challenge for recruiters in verifying the suitability of candidates’ qualifications for the required positions. Currently, the selection process still relies on Microsoft Excel as a tool for keyword searches in candidate data, requiring approximately 30 s to process a single keyword. The implementation of the Content-Based Filtering method in the applicant recommendation system is highly suitable for enhancing the recruitment process efficiency, as it focuses on analyzing the similarity of descriptions in the curriculum vitae data uploaded by candidates. This system operates by calculating the similarity of candidate data based on the entered keywords. The system development process includes several stages, such as data preprocessing, Term Frequency–Inverse Document Frequency (TF-IDF) calculation, Cosine Similarity computation, and ranking the results based on Cosine Similarity values from highest to lowest. The experiment results show that searching for a single keyword takes approximately 0.7 s, which is 29.3 s faster than the current candidate selection process. Additionally, a user satisfaction survey evaluated using the DeLone and McLean model, achieved a score of 95.4%, indicating that the system effectively enhances the accuracy and efficiency of applicant recommendations in the recruitment process.