An Innovative Approach to Predicting Employee Attrition through Optimized Neural Mining and Advanced Classification Techniques

Abstract

Employee turnover is a slow process of staff loss because of reasons such as retirement or personal choices, and it is not always replaced with a like in a similar time frame. In the context of organizations, appropriate forecasting of attrition is essential to ensure that the organizations retain talented personnel to aid in strategic planning of the workforce. The current paper proposes a new deep learning model, Optimized Neural Mining and Classification Logic (ONMCL) to enhance attrition prediction of employees. ONMCL combines state-of-art in neural network potentials and methodical pre processing of data, which could help unearth patterns in workers conduct and turnover proclivities. Traditional Logistic Regression (LR) is used to model it against so as to verify performance. Whereas LR had an accuracy of 93% in training, ONMCL performed better with accuracy more than 96%. Main drivers of attrition and their interdependencies are also discussed in the study. Findings demonstrate that ONMCL not only increases the accuracy of prediction but also offers organizations greater predictive insights relative to the causes of attrition with respect to the provision of more stable and actionable resource in respect of proactive advance planning retention of employees within organizations.