Comparative Study of Various Machine Learning and Deep Learning Techniques for Energy Prediction and Consumption Using IoT Modules

Abstract

Due to the rise in worldwide energy consumption, prediction of energy in the buildings which has recently become a critical subject of cost effectiveness and energy conservation (Almalaq and Zhang in IEEE Access 7:1520–1531, 2018 [1]). The insecurity of energy supply poses a threat to the systems’ energy efficiency and long-term viability. To plan the use of the captured energy, well-defined power management systems are required. Because of the complexities of the demand patterns that are available, predicting future electricity demand is a difficult task (Kang et al. in Appl Sci MDPI J 10:7241, 2020 [2]). Power management modules are used by certain intelligent IoT node networks to identify the frequency and voltage of each component of their hardware in actual environments. As per the predicted energy harvest. End-users rely on businesses that supply commercialized electrical energy in steady mode and safe electricity. As a result, developing effective models for prediction will be a critical stage to plan electronic power system operations (Kang et al. in Appl Sci MDPI J 10:7241, 2020 [2]). Because of its relevance in decreasing energy waste, the energy management system for building (BEMS) have become a topic of hot nowadays. However, due to issues such as low forecast accuracy. Energy usage forecasting is one of the application of BEMS’ which is remained as static (Shapi et al. in Dev Built Environ J 5:100037, 2021 [3]). Serious problem about lowering the energy usage in building requires using the effective prediction mechanism for correctly estimating energy in the future consumption (Almalaq and Zhang in IEEE Access 7:1520–1531, 2018 [1]). Here in the research article, we have performed a detailed comparative study of various machine learning and deep learning techniques for energy prediction and consumption using IoT modules and their performance with real-time data sets and historical-based data sets.