<![CDATA[Cyberloafing refers to the practice of employees utilizing internet access for non-job-related activities during work hours. Cyberloafing poses a dilemma for organizations, as it is deemed aberrant conduct that might impact overall performance. Consequently, organizations must ascertain the determinants of cyberloafing. This study seeks to identify a suitable predictive model for the determinants of cyberloafing behavior in the workplace using a machine learning methodology. The employed methodology utilizes the conventional data mining cycle, namely the Cross-Industry Standard Process for Data Mining (CRISP-DM), with Orange Data Mining as the application tool. The findings indicate that Logistic Regression is the most effective model for forecasting cyberloafing. Logistic Regression yields performance scores of 90.5% Precision and 88.9% Recall. Conversely, the Naïve Bayes model had the lowest metrics, with a Precision of 64.8% and a Recall of 51.9%. This study serves as a reference demonstrating that Logistic Regression effectively predicts cyberloafing. This study enables firms to examine the factors contributing to cyberloafing, facilitating the development of policies aimed at mitigating its adverse effects.]]>
Copyrights © 2026
