Forecasting the probability of solar power output using logistic regression algorithm

• Published in: Journal of statistics & management systems Vol. 23; no. 1; pp. 1 - 16
• Main Authors: Jagadeesh, V., Venkata Subbaiah, K., Varanasi, Jyothi
• Format: Journal Article
• Language: English
• Published: New Delhi Taylor & Francis 02-01-2020; Taru Publications
• Subjects: Algorithms; Datasets; Economic forecasting; Forecasting; Logistic Regression; Machine learning; Natural resources; Probability; Regression; Regression analysis; Solar energy; Statistical analysis; Sustainable development; Urbanization
• ISSN: 0972-0510; 2169-0014
• DOI: 10.1080/09720510.2020.1714146

Now days, Natural resources are depleting due to rapid urbanization and increasing demand of energy requirement. Hence, Renewable energy plays a vital role in sustainable development. The role and selection of solar variables and applying machine learning algorithm plays a vital role for forecasting the probability of solar power output for any selected day in the future. This article Includes datasets collected for over a period for around eleven months output of plant, incident solar radiations, and local temperature. These datasets are feed to one of supervised machine learning algorithms with logistic regression for short term forecasting of solar plant output on a particular day in future using previous year data. For efficient forecasting of solar plant output proper variable are to be selected and used to feed the logistic regression algorithm to obtain efficiency in training the data and predict the plant output. In this article the accuracy obtained for amount of data feed is studied and also the forecast the probability in order to predict the output from that plant.