<![CDATA[This research intends to examine the satisfaction level of State Civil Apparatus (ASN) employees in the South Bangka Regency Administrator regarding the implementation of the BKN E- Kinerja System by utilizing Machine Learning Technology classification technique, specifically the Decision Tree and Naïve Bayes algorithms. The data from the research were gathered via surveys administered to ASN personnel utilizing Google Form featuring a Likert Scale and categories of classification established by the researcher. The variables utilized in this research encompass perceived utility, system performance and perceived user-friendliness. Data processing was conducted using the Kaggle. Platform via multiple phases, encompassing data preprocessing, division of training and testing sets, classification processes and model assessment utilizing confusion matrixm accuracy, precision, recall and f1- score. The outcome of the test shows that the Naïve Bayes Algorithm performed better tha the Decision Tree Algorithm achieved an accuracy of 84.39% while the Decision Tree reached an accuracy a value of 82.44%. In the Decision Tree model, the confusion matrix showed that 149 data instances were correctly classified in class 1 and 20 data instances in class 0. Feature importance analysis revealed that the perceived usefulness variable was the most significant factor with an importance value of 0.571632, followed through system quality and perceived ease of use. From these findings, it can be inferred that the Naïve Bayes the algorithm is more efficient for categorizing user satisfaction levels in this research dataset as it generates greater precision in comparison to the Decision Tree Algorithm. This research is anticipated to function as a assessment and the evaluation of resources in enhancing the service quality of the BKN E- Kinerja System.]]>
Copyrights © 2026
