Department of Mechanical Engineering, SRM Institute of Science and Technology, Ramapuram, Tamilnadu, India.
Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, 602105, India.
Chemosphere. 2022 Dec;308(Pt 1):136277. doi: 10.1016/j.chemosphere.2022.136277. Epub 2022 Sep 1.
The consumption of a significant quantity of energy in buildings has been linked to the emergence of environmental problems that can have unfavourable effects on people. The prediction of energy consumption is widely regarded as an effective method for the conservation of energy and the improvement of decision-making processes for the purpose of lowering energy use. When it comes to the generation of positive results in prediction tasks, the Machine Learning (ML) technique can be considered the most appropriate and applicable strategy. This article presents a Modified Wild Horse Optimization with Deep Learning approach for Energy Consumption Prediction (MWHODL-ECP) model in residential buildings. The MWHODL-ECP method that has been provided places an emphasis on providing an up-to-date and precise forecast of the amount of energy that residential buildings consume. The MWHODL-ECP algorithm goes through several phases of data preprocessing in order to achieve this goal. These steps include merging and cleaning the data, converting and normalising the data, and converting the data. A model known as deep belief network (DBN) is used here for the purpose of predicting energy consumption. In the end, the MWHO algorithm is utilised for the hyperparameter tuning procedure. The results of the experiments demonstrated that the MWHODL-ECP approach is superior to other existing DL models in terms of its performance. The MWHODL-ECP model has improved its performance, with effective prediction results of MSE-1.10, RMSE-1.05, MAE-0.41, R-squared-96.28, and Training time-1.23.
建筑物的大量能源消耗与环境问题的出现有关，这些问题可能对人们产生不利影响。能源消耗预测被广泛认为是节约能源和改进决策过程以降低能源使用的有效方法。在预测任务中产生积极结果方面，机器学习 (ML) 技术可以被认为是最合适和适用的策略。本文提出了一种用于住宅建筑能耗预测的改进野马优化与深度学习方法 (MWHODL-ECP) 模型。所提供的 MWHODL-ECP 方法侧重于提供住宅建筑能耗的最新和精确预测。MWHODL-ECP 算法通过几个数据预处理阶段来实现这一目标。这些步骤包括合并和清理数据、转换和归一化数据以及转换数据。这里使用深度置信网络 (DBN) 模型来进行能耗预测。最后，MWHO 算法用于超参数调整过程。实验结果表明，MWHODL-ECP 方法在性能方面优于其他现有的深度学习模型。MWHODL-ECP 模型通过提高其性能，实现了有效的预测结果，MSE 为 1.10，RMSE 为 1.05，MAE 为 0.41，R-squared 为 96.28，训练时间为 1.23。
