Modelling Willingness to Pay of Electricity Supply Using Machine Learning Approach

Abstract

Background: Electricity has long been regarded as a catalyst for enhancing economy growth in developing countries. The availability of electricity can drive a country’s growth by allowing firms to take advantage of it to increase its productivity enhancing technologies, the bulk of which are reliant on electricity. Objective: In view of the above, this study aims to model the willingness to pay of electricity supply to consumers in southwestern states of Nigeria using machine learning approach. Methods: The study was conducted in six southwestern states in Nigeria. This data contain information which was obtained through longitudinal survey using a google form questionnaire. The data contains the category of the respondents, type of commercial activity, average daily electricity supplied, willingness to pay of electricity supply per month and so on. K-Nearest Neighbors (KNN), Random Forest, Support Vector Machine (SVM), Decision Tree and Boosting Classification models were considered. Results: Among the models considered, KNN, SVM and Boosting classification models perform better in classifying whether a consumer is willing to pay for electricity consumption or not. Conclusion: The information obtained from this research can be used to produce insight into electricity production in the states.