Whistleblowing is a complaint system and follow-up management of each violation report. The problem that arises is when determining the follow-up, namely determining the severity or severity of the violation and the sanctions given are only based on the superior's assessment without adhering to standard guidelines or rules. This results in the sanctions given not in accordance with the violations committed. The purpose of this study is to classify the types of violations so as to facilitate the determination of sanctions on the whistleblowing system using the C4.5 Algorithm. The partition was performed three times with the highest additional value of 0.8516 and a decision tree was obtained. Based on the decision tree, the final node that has been generated is then extracted into 27 rules. The classification results from the C4.5 Algorithm can be used to classify the types of violations with an accuracy rate of more than 80%. The first validation with 15 tests obtained an accuracy rate of 86.66%. The second validation is the combination of data on 125 cases of combination data and obtained an accuracy rate of 84.8%. The decision tree generated from three partitions consists of 27 rules that can be used as a pattern to classify the types of violations.
Copyrights © 2023