An Ensemble Learner-Based Bagging Model Using Past Output Data for Photovoltaic Forecasting

Abstract

As the world is aware, the trend of generating energy sources has been changing from conventional fossil fuels to sustainable energy. In order to reduce greenhouse gas emissions, the ratio of renewable energy sources should be increased, and solar and wind power, typically, are driving this energy change. However, renewable energy sources highly depend on weather conditions and have intermittent generation characteristics, thus embedding uncertainty and variability. As a result, it can cause variability and uncertainty in the power system, and accurate prediction of renewable energy output is essential to address this. To solve this issue, much research has studied prediction models, and machine learning is one of the typical methods. In this paper, we used a bagging model to predict solar energy output. Bagging generally uses a decision tree as a base learner. However, to improve forecasting accuracy, we proposed a bagging model using an ensemble model as a base learner and adding past output data as new features. We set base learners as ensemble models, such as random forest, XGBoost, and LightGBMs. Also, we used past output data as new features. Results showed that the ensemble learner-based bagging model using past data features performed more accurately than the bagging model using a single model learner with default features.

PAN-LDA: A latent Dirichlet allocation based novel feature extraction model for COVID-19 data using machine learning

“…Y. Tounsi, L. Hassouni, and H. Anoun have introduced a new model CSMAS to predict problems in data mining of credit scoring domain using state-of-the-art gradient boosting methods (XGBoost, CatBoost, and LightGBM) [ 25 ]. Sunghyeon Choi forecasted solar energy output by employing RF, XGBoost, and LightGBM models [ 26 ]. Moreover, J.Cordeiro, O. Postolache, and J. Ferreira used the XGBoost model and the LightGBM model to predict the height of children [ 27 ].…”

“… Reference Year Base model Time series prediction Data set [ 28 ] 2021 SVM, LR, Multi-layer perceptron, RF Covid-19 pandemic cumulative case forecasting Data of COVID-19 between January 20, 2020, and September 18, 2020, for the USA, Germany, and global was obtained from the World Health Organization website. [ 17 ] 2020 LR, SVM, ANN, RF, XGBoost, LightGBM Prediction of peak demand days of cardiovascular disease(CVD) admissions Health Information Center of Sichuan Province, China: the daily number of admissions of CVD patients in hospital Chengdu Meteorological Monitoring Database: Meteorological data China National Environmental Monitoring Cente: air pollutants data [ 25 ] 2020 XGBoost, CatBoost, LightGBM Prediction of problems in data mining of credit scoring domain Home Credit Default Risk from Kaggle Challenge [ 26 ] 2020 RF, XGBoost, LightGBM Photovoltaic Forecasting Data of a Photovoltaic plant in South Korea [ 18 ] 2020 LightGBM Cryptocurrency price trend Daily trading data from https://www.investing.com/ [ 29 ] 2020 XGBoost, ARIMA Hemorrhagic fever with renal syndrome Monthly hemorrhagic fever with renal syndrome incidence data from 2004 to 2018 from the official website of the National Health Commission of the People's Republic of China [ 27 ] …”

Forecast Quality of Physics-Based and Data-Driven PV Performance Models for a Small-Scale PV System

“…When basic learners of the same type are used, the approach is referred as homogeneous ensemble model and when the basic learners are built of different algorithms, it is called heterogeneous ensemble method. • Bagging (BAG) (Choi and Hur, 2020) regressors help to improve model performance by training in parallel each basic learner on a random subset of the training dataset and averaging the resulting single forecasts. • Boosting (XGB [for Extreme Gradient Boosting]) is like bagging but runs sequentially.…”

Prediction with Multi-layer Perceptrons of Greenhouse Gas Emission Belonging to Different Sectors in Turkey

“…Çeşitli çalışmaların sonuçları, geçmiş veri özelliklerini kullanan topluluk öğrenme tabanlı modelin, varsayılan özelliklere sahip tek bir model kullanarak gerçekleştirilenden daha doğru performans sağladığını göstermiştir (Choi & Hur, 2020;Li, Han, Wang, & Zhao, 2018). Topluluk yöntemleri kullanılarak CO2 emisyonunun değerinin tahmin edilmesi konusunda çeşitli araştırmalar yapılmıştır (Choi & Hur, 2020;Parker, 2010). Khan & Awasthi (2020), çalışmalarında Kanada'daki yolcu ve yük karayolu taşımacılığından kaynaklanan sera gazı emisyonlarını tahmin etmek için veri madenciliği ve denetimli makine öğrenme algoritmalarına (regresyon ve sınıflandırma) dayalı yeni modeller önermişlerdir.…”

“…Topluluk öğrenmesinin amacı, tek bir sınıflandırıcı üzerindeki öngörüyü genelleştirerek doğruluğu arttırmada çeşitli temel sınıflandırıcıların kararlarını veya tahminlerini birleştirmektir (Chen, Zahiri, & Zhang, 2017). Çeşitli çalışmaların sonuçları, geçmiş veri özelliklerini kullanan topluluk öğrenme tabanlı modelin, varsayılan özelliklere sahip tek bir model kullanarak gerçekleştirilenden daha doğru performans sağladığını göstermiştir (Choi & Hur, 2020;Li, Han, Wang, & Zhao, 2018). Topluluk yöntemleri kullanılarak CO2 emisyonunun değerinin tahmin edilmesi konusunda çeşitli araştırmalar yapılmıştır (Choi & Hur, 2020;Parker, 2010).…”