Predictive Talent Acquisition Using Ensemble Learning Models in Human Resource Management

Abstract

Predictive talent acquisition is one of the applications of machine learning which has become an area of importance within human resource management for carrying out data analysis of recruitment processes. The purpose of this research was to compare the effectiveness of several machine learning classifiers on the prediction of recruitment decision. For that, a dataset containing structured information about 1,500 candidates was used. Logistic Regression, Decision Tree, Random Forest, Extra Trees, and HistGradientBoosting were the five classifiers considered in the study. These classifiers were evaluated using a stratified train/test split of 80:20. Their performances were measured in terms of accuracy, precision, recall, F1-score, area under the ROC curve (ROC-AUC), area under the PR curve (PR-AUC), and Brier score. The best result was demonstrated by HistGradientBoosting classifier with accuracy 93.67%, F1-score 0.8914, ROC-AUC 0.9313, and PR-AUC 0.9173. The feature importance analysis yielded the recruitment strategy, skill score, and educational background as some important predictive factors, although it was necessary to consider the problem of target leakage in relation to the recruitment strategy. The outcomes prove that the prediction capability is good enough for the experimental dataset but at the same time emphasize the significance of validation, correct use of features, fairness analysis, and human supervision. Since the dataset in question is artificial, the outcomes cannot be regarded as an indication of effectiveness in the real-world recruitment process.