|
Abstract:
|
Access to electricity is essential to life in the modern society and is increasingly becoming a basic human right. To meet these increasing demands for energy across the globe and to offset the detrimental effects that burning fossil fuels have on the environment, researchers have been investigating the use of renewable energy resources, such as solar energy. Accurate solar resource data are essential for viability assessments of potential solar energy installations. Outliers in such data can significantly impact the accuracy of these assessments. Methods currently used to detect outliers in solar resource data do not adequately identify these outliers. This study investigates and compares the use of traditional outlier detection methods to several classification methods, including kNN, naïve Bayes and tree-based methods, for the purpose of outlier detection in this field. The results show that the tree-based methods provide accurate identification of outliers and are demonstrated on a data set collected in Gqeberha, South Africa. The use of the proposed approach can aid in reducing the uncertainty in solar resource data and hence promote the use of solar energy solutions.
|
