The Application of Data Mining Techniques for Analyzing Electricity Usage Behavior of Consumers in Thailand

Abstract

This study aims to investigate the relationship between electricity user types and their rank of monthly electricity consumption, analyze the electricity usage patterns using the Apriori algorithm, and propose energy policies based on the analysis results. The sample data comprises 3,562 records of electricity consumption from a government database, covering residential, general business, large-scale industry, backup electricity users, and specialized business types. This approach differs from other research which often focuses on smart meter or prepaid bill data. The research technique employed is data mining using the Apriori algorithm, which has an advantage over Clustering or Prediction methods as it can discover association rules without requiring a pre-defined number of groups. The statistical measures used are confidence and association (lift). The findings show that large-scale industries and businesses have the highest electricity consumption volume every month, while backup electricity users have the lowest usage. Residential users consuming less than 150 units and general businesses fall into the middle rank. Furthermore, the Apriori algorithm demonstrated high accuracy in discovering significant relationships, which can be used to develop policies, such as optimizing energy resource allocation for industries, restructuring backup electricity tariffs, and promoting efficient electricity consumption in residential and general business sectors to create a sustainable energy system, requiring significant results such as the relationship between industry and high electricity usage, can be practically applied in planning to increase production capacity during high-demand periods.