Comparative Study of C45 and SVM Machine Learning Classification in Predicting Electrical Supply Disturbances

Authors : Christofer Satria; Fanar Adi Apriliawan; Anthony Anggrawan; Mayadi Mayadi; Peter Wijaya Sugijanto et al.
article cite 2 Year 2024
source:
Abstract

Disturbances in the electricity supply disrupt people's daily activities and increase community operational costs. In other words, power outages or supply disruptions negatively impact many parties and need to be resolved. Analysis of the causes of disruption to the customer's electricity supply/network is required. Artificial intelligence (AI) based application is one approach that can predict the cause of power outages. Machine learning (ML) is a part of artificial intelligence and the primary key to intelligently predicting based on class feature categories. Bearing in mind that analytical errors often occur in classification or prediction based on the available data set, of course, this is a challenge to solve. Fortunately, ML has become a powerful tool that helps classify or predict effectively and better. That is why this research aims to predict or classify disturbances that cause electricity supplies to break or go out to customers. The prediction method used in this research is the C4.5 learning machine and Support Vector Machine (SVM). The research results on the classification of types of electrical disturbances show that ML C4.5 has an accuracy of 84.21%, precision of 66.67%, and recall of 100.00%. Meanwhile, ML SVM produces an accuracy of 89.47%, a precision of 77.78%, and a recall of 100.00%. This means that ML SVM predicts the cause of electricity supply disruptions more accurately than ML C4.5.


Concepts :
Machine Fault Diagnosis Techniques
article cite 2 Year 2024 source
Citations by Year
YearCount
2024 2