Enhancing Organizational Performance Through KPI-Based Employee Prediction Using the C4.5 and Random Forest Methods

Abstrak

Employee performance evaluation is a crucial process for organizations in achieving strategic goals. PT Kreatif Dinamika Integrasi currently conducts employee assessments manually using Microsoft Excel, leading to inefficiencies, potential errors, and subjectivity. This study aims to develop a decision support system (DSS) using the C4.5 and Random Forest algorithms to improve accuracy and fairness in performance evaluation. The research adopts a quantitative approach, encompassing problem analysis, literature review, data collection, machine learning model implementation, and result evaluation. The key performance indicators (KPIs) used include discipline, punctuality, skills, appearance, and education, which serve as attributes in the classification process. The C4.5 algorithm constructs a decision tree to identify patterns, while Random Forest, as an ensemble method, enhances classification accuracy and reduces the risk of overfitting. The results indicate that using Random Forest improves evaluation accuracy from 85% to 87.5%. The implementation of this DSS provides a more reliable framework for salary and bonus prediction, minimizes bias, and enhances decision-making quality. Overall, integrating machine learning models into employee performance evaluation significantly improves efficiency and transparency.