Comparison of Machine Learning Prediction Models on Domain-specific Recruitment Data

Abstract

A new domain that Artificial intelligence (AI) has impacted is downstream tasks of talent management in organizations. Though human agents are unavoidable in the recruitment process, resume classification and ranking tasks could be greatly assisted by intelligent modeling techniques for the classification and ranking of applicants’ skill sets. In this study, machine learning (ML) and natural language processing (NLP) were used for resume classification and ranking. A secondary dataset that focused on domain-specific fields of information technology job applicants’ resumes was collected and curated. The dataset was modeled using four machine learning algorithms such as k-nearest neighbor (KNN), Random Forest (RF) classifier, support vector machine (SVM), and Decision tree (DT), and the cosine similarity and TF-IDF Vectorizer were used to find the resume that is the most comparable to the job description provided in the job adverts. Experimental results showed that the decision tree had the highest performance of 99%, followed by random forest 98%, support vector machine 97%, and k-nearest neighbor 80%. The study found that high-performing models could greatly impact resume classification and prediction. More so, AI technologies have significant impacts on the recruitment and Selection Procedure with positive outcomes in accuracy, time, cost, and fairness.