ARCHIVES

Due to the large number of resumes obtained for each job posting, homemade screening resumes have become inefficient, time-consuming, and biased with the growing number of online recruiting sites. We address in this paper the problem of resume screening with a machine learning approach and examine the ability some learning algorithms have in correctly predicting how good a candidate fits to the given job description. In this paper, we focus on the transformation of raw unstructured resume information to structured form related to text cleaning, tokenization, stop word removal and a few other feature extraction methods such as Term Frequency–Inverse Document Frequency (TF-IDF) and word embeddings.[1][2] In the paper, we discuss and compare various methods used for identifying job relevancy, such as logistic regression, support vector machines (SVM), and random forest based on job titles. For the purpose of testing the accuracy and robustness of the models, global quality indicators accuracy, precision, recall, and F1-score are adopted. Perform a comparative study of the two methods in terms of the pros and cons of the two. Our research, and its outcomes, demonstrates that the proposed machine learning-based methods are not only effective to improve the efficiency and fairness of resume screening but also the best model for the given scenario could be determined in an automated recruitment system. This paper advances the state-of-the-art for smart talent acquisition solutions by providing insights on the selection of models for resume screening tools. Furthermore, this research incorporates real dataset-based experimentation and advanced visualization techniques such as confusion matrices, ROC curves, and cross-validation analysis to provide deeper insights into model performance and reliability.