Enhancing Proactive Talent Management in Private Banking with Machine Learning-Based Evaluation of Person-Organization Fit

In this research, aim is to propose a data-driven approach for enhancing talent management strategies in the private banking sector by evaluating person-organization fit dynamics. To achieve this aim, OSCN and RNN with the attention mechanism by analyzing a dataset of 120 employees profiles, performance evaluations, turnover rates, and organizational structures. The weighted importance, as well as the results of the OSCN and the attention model, are presented in the paper. The assessment of the quality of predictions in terms of turnover has demonstrated their high accuracy; the average indices are as follows: 82% for the accuracy of predictions, 85% for precision, 78% for recall, 81% for the F1 score, 0.87 for the AUC-ROC, and 0.79 for the AUC-PR. In the context of job performance predictions, the results for MSE, RMSE, and MAE constitute low values of 0.025, 0.158, and 0.124, respectively. It is equally important to note that the models have also demonstrated a high level of interpretability. Represented through the R 2 metrics, it can account for the 60% total variance for the ANN, though the SVM is capable of predicting up to 40% of the future performance level. Furthermore, the Spearman’s rank correlation coefficient is also relatively high and demonstrates a strong relationship between the actual and predicted performance levels.