ICFICE 2019
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Comparison of Machine Learning Models for Solar Energy Prediction |
¿µ¹®Á¦¸ñ(English Title) |
Comparison of Machine Learning Models for Solar Energy Prediction |
ÀúÀÚ(Author) |
Wonseok Jung
Young-Hwa Jeong
Moon-Ghu Park
Jeongwook Seo
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 11 NO. 01 PP. 0341 ~ 0344 (2019. 06) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
In this paper, we compare the performance of Support Vector Regression (SVR) and Artificial Neural Network (ANN), which are machine learning models using actual measured weather data, to present the best predictive model for arbitrary solar panels in a specific region. In the solar panels each installed in each of the 15 different regions, actual measured seven types of weather data such as solar radiation, temperature, precipitation, daylight hours, snowfall, cloudy, and visibility are used to generate three prediction models. Then the parameters of the prediction model suitable for a solar panel in a specific region are derived through experimental studies. Mean Absolute Percentage Error (MAPE) and Coefficient of Variance (CV) were used to compare the performance of the generated three prediction models. According to the result of the experiment of the solar panel (2) of the specific region (133), the SVM model achieved the performances of MAPE = 17.91% and CV = 4.48%, and the ANN model MAPE = 19.95% and CV = 5.68%. In addition, a prediction model suitable for each region is suggested by comparing the performance of the prediction models developed to derive the best prediction models for each region.
|
Å°¿öµå(Keyword) |
Solar Energy
Prediction
Support Vector Machine
Artificial Neural Network
Weather data
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|