EUVIMEDEUVIMED entdecken
Crossref

Comparative Performance Analysis of Solar Energy Systems using Linear Regression and Support Vector Regression

R. Saravanan

REST Journal on Data Analytics and Artificial Intelligence · 01.03.2024 · 3 (1) · 82

DOI: 10.46632/jdaai/3/1/10 ↗

Kurz gefasst

Fachartikel in der Zeitschrift REST Journal on Data Analytics and Artificial Intelligence, online erschienen am 01.03.2024. Verfasst von R. Saravanan. Die Arbeit nennt 29 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The necessity for sustainability in infrastructure has resulted in the rapid adoption of solar energy systems in building designs. The study aimed to create predictive modeling for the output energy generated in solar energy systems through the application of machine learning models, namely Linear Regression and Support Vector Regression, based on the given dataset with 20 parameters and corresponding parameters such as solar irradiance, panel efficiency, panel area, and ambient temperature, among others. The results from the descriptive statistical analysis showed that the average solar irradiance was 5.65 kWh/m²/day, while the average panel efficiency was 19.28%. The energy output was averaged to be 27.33 kWh/day, while nearly perfect product correlation was established for each input variable and energy output, as shown by r = 0.98 and r = 0.99 for each corresponding value. The Linear Regression model had a high R² value, which was 0.9948 for the training set and 0.9767 for the test set, thus proving the efficiency of the model in generalizing the data with minimal over-fitting. The Support Vector Regression model had an extremely high R² value, which was 0.9973 for the training set, and a high R² value, which was 0.9505 for the test set, thus validating the efficiency of this model as well. It can be noted that both models had extremely high accuracy in predicting the energy output based on the panel area. Additionally, it can be noted that both models had extremely low error values for all parameters, namely RMSE, MAE, and MSLE. This further validates the fact that machine learning models like LR and SVR are highly efficient in the prediction of solar energy output. This research can be said to be contributing to the overall objective of incorporating data science techniques into the design of Building Integrated Photovoltaic (BIPV) systems.

Zitieren

APA

Saravanan, R. (2024). Comparative Performance Analysis of Solar Energy Systems using Linear Regression and Support Vector Regression. REST Journal on Data Analytics and Artificial Intelligence, 3(1), 82. https://doi.org/10.46632/jdaai/3/1/10

Vancouver

Saravanan R. Comparative Performance Analysis of Solar Energy Systems using Linear Regression and Support Vector Regression. REST Journal on Data Analytics and Artificial Intelligence. 2024;3(1):82. doi:10.46632/jdaai/3/1/10

BibTeX
@article{Saravanan2024,
  author = {Saravanan, R.},
  title = {Comparative Performance Analysis of Solar Energy Systems using Linear Regression and Support Vector Regression},
  journal = {REST Journal on Data Analytics and Artificial Intelligence},
  year = {2024},
  volume = {3},
  number = {1},
  pages = {82},
  doi = {10.46632/jdaai/3/1/10}
}
Kostenlos starten →