In the recruitment process, hand-picking the right candidate out of a pool of resumes in the allotted time can be very challenging. Furthermore, as there is no standard pattern for producing CVs, there is always the risk of overlooking crucial information. This research focuses on extracting important information from CVs and job descriptions of various formats, using several machine learning techniques. The extracted data from CVs and job descriptions is then compared to find the best-matched CVs using a variety of metrics such as cosine similarity, soft cosine similarity, Jaccard similarity, dice similarity coefficient, overlap coefficient, and conditional probability. Conditional probability outperforms other metrics and hence is chosen as the metric to assess similarity. Following this, the content-based recommendation technique is used to recommend candidates based on skillset. Currently, our work is limited to computer science, but we plan to extend our work into other sectors in the future.
CITATION STYLE
Shovon, S. M. S. F., Mohsin, M. M. A. B., Tama, K. T. J., Ferdaous, J., & Momen, S. (2023). CVR: An Automated CV Recommender System Using Machine Learning Techniques. In Lecture Notes in Networks and Systems (Vol. 597 LNNS, pp. 312–325). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-21438-7_25
Mendeley helps you to discover research relevant for your work.