A DATA MINING APPROACH FOR DETERMINING POWER CONSUMPTION OF NIGERIAN ELECTRIC UTILITY CUSTOMERS
ABSTRACT
The electric industry plays a pivotal role as a primary service provider and serves as the backbone of the global energy sector. In Nigeria, the national electric utility stands as the sole entity responsible for distributing electrical power throughout the nation. The electric power sector is marked by its dynamic and highly competitive nature, necessitating rapid adaptation to meet the evolving needs and desires of both individual and organizational customers. According to Energy pedia’s 2016 publication, merely 27% of Nigeria’s population has access to the electricity grid. In light of this, the objective of this research is to develop a predictive model for estimating power consumption among Nigeria’s electric utility customers by employing data mining techniques. This study draws from data collected from Nigeria’s electric utility customers to harness big data for analysis. The research employs a hybrid data mining methodology, which encompasses customer classification based on power consumption patterns and the construction of a prediction model using classification algorithms. The research process involves several key stages, including problem identification, data comprehension, data preparation, modeling, knowledge discovery evaluation, and the design of a user interface to facilitate the application of the newfound insights. The dataset spans from January 2008 to January 2011 E.C. and encompasses data from all Nigeria’sutility customers, comprising 14 attributes with a total of 85,849 instances. Four classification algorithms were utilized to construct predictive models: J48, bagging, random tree, and PART. The experimental results indicate that the J48 algorithm outperformed the other models, achieving an accuracy rate of 96.61%. Specifically, the J48 model accurately classified 82,939 instances (96.61%), with only 2,910 instances (3.38%) being incorrectly classified. The primary outcome of this study is the classification of power consumption predictions for new connections of electric utility customers, distinguishing between high and low power consumption patterns. Based on the research findings, the researcher recommends that the electric industry explore further studies to develop a system enabling optimal management of power consumption.
A DATA MINING APPROACH FOR DETERMINING POWER CONSUMPTION OF NIGERIAN ELECTRIC UTILITY CUSTOMERS, GET MORE COMPUTER SCIENCE PROJECT TOPICS AND MATERIALS