# Application of Support Vector Machine (SVM) Method for Photovoltaic Condition Classification Based on Characteristic Curve Indicators
> Assalam I.F.

URL kanonis: https://discover.unhas.ac.id/publications/pub_scopus_86000017999
Jurnal / Konferensi: Proceeding 2023 International Conference on Artificial Intelligence Robotics Signal and Image Processing Airosip 2023
Tahun terbit: 2023
DOI: https://doi.org/10.1109/AIRoSIP58759.2023.10873920
Citations: 0

## Authors
- Assalam I.F.

## Abstract
The percentage of degradation rate is a major factor in power reliability indicators in photovoltaic. To evaluate the reliability, a condition analysis is performed which includes hotspot, bypass diode failure, and short circuit by measuring the characteristic curve. This research aims to apply the Support Vector Machine (SVM) method in applying a multi-classification process that will be validated and combined with the Naïve Bayes (NB) and K-Nearest Neighbors (KNN) methods. In the analysis conducted, various SVM methods with kernel variations such as Linear SVM, Polynomial SVM, and Gaussian SVM were evaluated and combined with NB and KNN. The results showed that the Polynomial SVM-KNN, Gaussian SVM-KNN, and KNN methods provided excellent accuracy rates in the multi-classification process based on the specified data while the combination with NB showed that it did not provide significant advantages in improving the accuracy rates. In this case, the Polynomial SVM-KNN method achieved an accuracy rate of 98.33% train and 87.00% test, while the Gaussian SVM-KNN method achieved an accuracy rate of 97.62% train and 88.00% test. In addition, the KNN method also provides a very high level of accuracy with a train value of 98.33% and a test of 92.00%. This research provides recommendations for using Polynomial SVM-KNN, Gaussian SVM-KNN, or KNN methods in the classification of photovoltaic conditions with a high level of accuracy.

## Keywords
- Support vector machine
- Photovoltaic system
- Computer science
- Pattern recognition (psychology)
- Artificial intelligence
- Relevance vector machine
- Machine learning
- Engineering
- Electrical engineering

---
Sumber: Discover Unhas — RIMS Universitas Hasanuddin.
Saat mengutip, gunakan DOI bila tersedia atau URL kanonis di atas.