Comparison of Classification Algorithms on Household Electricity Consumption Data

Abstract

The pattern of electricity consumption is one thing that is important to be known by a household, so it is essential to identify the type of intensity of electricity usage from the household's daily life. It can help determine how much electricity consumption of equipment so that efforts can be made to optimize electricity consumption further while saving costs. Due to that, the classification algorithms based on supervised learning is used. In this study, we compared several types of classification methods to determine the type of electricity usage patterns in a daily household life on Household Electric Power Consumption data obtained from Kaggle. The classification methods being compared are KNN, SVM, Decision Tree, and Logistic Regression. The accuracy of all methods is analyzed to find which method is best in identifying the intensity of electricity usage. From the results of this study, it was found that the Logistic Regression method was the most accurate in classify ing the type of intensity of electricity consumption with an average accuracy value of 99%.