Towards Efficient Energy Management of Smart Buildings: A LSTM-AE Based Model

Optimal home energy utilization and the development of residential segment renewable energy systems are two problems that might be addressed by interacting energy systems in smart homes. Reduced grid electricity usage, particularly during peak hours, increased economic efficiency, user satisfaction (via power seller/purchaser choice), and consumer engagement in energy management are the primary goals of these energy interactions. Performing preprocessing, feature selection, and model training in the proper order is of the utmost importance. Using these measures for data preparation can lead to erroneous model interpretations or lower model performance. Principal component analysis (PCA) is a well-liked method for feature selection that uses linear transformation to reduce the dimension of data in a high-dimensional feature space. When contrasted with the cutting-edge LSTM and AE, this method seems to be quite innovative. A considerable improvement in accuracy, reaching 97.50%, was demonstrated by the results.