PREDICTIVE MODELING OF HOUSEHOLD ENERGY CONSUMPTION IN THE USA: THE ROLE OF MACHINE LEARNING AND SOCIOECONOMIC FACTORS

Аннотация

INTRODUCTION that depict energy consumption levels and their Background and Motivation pattern. For example, larger households may have According to Hasan (2024), household energy larger homes and a proliferation of energy-using consumption encompasses a substantial portion of appliances. On the other hand, low-income total energy usage in the United States. According households may face several barriers, including to the United States Energy Information limited access to energy-efficient technologies that Administration, in recent years, residential energy contribute to higher energy inefficiencies. This use has made up about 22% of total energy calls for the need to recognize and address such consumption. This huge share underlines how disparities to achieve equity and efficiency in important households are in shaping the nation's energy access (Amiri et al., 2023; Chen et al, 2023). energy profile and environmental footprint. Problem Statement Understanding the pattern of energy use at the household level becomes ever more urgent in light Karmakar et al. (2024), posited that predicting of growing concerns about climate change and household energy consumption presents resource sustainability. Nasiruddin et al. (2023), noteworthy challenges because of the diverse reported that Energy use depends upon various array of factors that influence energy use, ranging factors, such as climate, household characteristics, from interacting physical attributes of the home, and behavior. Of these, income, education, and size like size and age, to behavioral elements of energy of the family are very vital socio-economic factors conservation practices to broader socioeconomic https://www.theamericanjournals.com/index.php/tajet 100 THE USA JOURNALS THE AMERICAN JOURNAL OF ENGINEERING AND TECHNOLOGY (ISSN – 2689-0984) VOLUME 06 ISSUE12 determinants. These relations are nonlinear, hence energy consumption. All this information on family demanding advanced levels of techniques to be income level, family size, and attainment of captured or modeled. It is in this aspect that education would help map changes in energy Machine Learning can provide promising usage behaviors. From such findings, one can solutions. In contrast to traditional statistical develop a practical and targeted energy efficiency approaches, Machine Learning models can handle program or policies as well as enable utilities to big, heterogeneous datasets and find complex propose strategies for reducing energy use and patterns in the data. Alam et al. (2023), asserted promoting sustainability. that while ML techniques have been applied to Significance of the Study perform well in different predictive tasks, The findings of this study have large implications incorporating socioeconomic factors into such for policy, the energy provider, and broader models remains largely unexplored. This is society in general. First is that the forecast of the considered a serious gap in the literature because consumption of energy at an individual household it reduces the understanding of how variables like level may underpin intervention designs targeted income and education contribute to disparities in at bringing about waste reduction and increased energy consumption, thus making the efficiency. For example, utilities may use forecast development of fair and efficient energy policies peaks to make dynamic pricing per time to difficult (Charfeddine et al., 2023; Goriparthi, motivate the shift of household usage to off-peak. 2023). Second, factoring in socioeconomic variables in Research Questions these predictive models allows the analysis of RQ1: How can machine learning models be patterns in energy consumption inequality at a used to predict household energy consumption much deeper level. Gaining this insight is crucial to in the USA? ensuring fairness and equal access to policies that make energy-efficient technologies and programs This research question aims to examine the available. For instance, identifying communities possibility of machine learning techniques to with the highest energy inefficiencies might guide predict household energy consumption. investments in energy retrofitting and education Techniques for machine learning can identify programs. This case scenario represents the latent patterns and trends by analyzing historical application of ML for energy prediction in the data on energy usage, alongside other related general direction of digital transformation and factors such as weather conditions, appliance smart grid technologies that will contribute to ownership, and occupancy. This can then be used more resilient and adaptive energy systems. With for prediction in forecasting future demand and these advantages, the United States will be able to may allow utilities to optimize resource allocation take big steps toward its goals of efficient and and grid management while offering personalized sustainable energy. advice to consumers on saving energy. Literat