<![CDATA[Population growth in urban areas drives a significant increase in household energy consumption. This condition poses a major challenge for the implementation of the smart city concept, particularly in achieving energy efficiency and sustainability. This study aims to classify household energy consumption patterns based on household power consumption data to support intelligent decision-making in urban energy management. The research method includes data preprocessing, data cleaning, and aggregation of daily energy consumption by utilizing key attributes such as Global Active Power, Voltage, Global Intensity, and three sub-metering variables. Consumption pattern categories are formed using the tertile method into three classes: Low, Medium, and High. Several machine learning algorithms are applied to build the classification model, including Logistic Regression, K-Nearest Neighbors (KNN), Random Forest, and Gradient Boosting. The test results show that the Random Forest model with hyperparameter adjustments produces the best performance with an accuracy value of 0.98 and an F1-macro value of 0.98, surpassing other models. These findings indicate that the ensemble learning approach is able to capture the complexity of household energy consumption patterns more effectively than conventional linear models. The contribution of this research lies in the development of a machine learning-based predictive model to support adaptive energy consumption monitoring and control systems in smart city implementations.]]>
Copyrights © 2025
