Harnessing Deep Learning for Enhanced Energy Consumption Forecasting in smart Home: A comparative Study of MLP and RNN Architectures

Each country is looking for effective ways of reducing energy consumption in residential sectors. The process needs close monitoring to comprehend energy usage trends. It is, therefore, important to predict energy consumption in households that will help in formulating efficient ways of optimization...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2025 3rd International Conference on Electronics, Energy and Measurement (IC2EM) s. 1 - 5
Hlavní autori: Sarah, Younsi, Rabea, Guedouani, Amirouche, Nait Seghir
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 06.05.2025
Predmet:
Accuracy
Deep learning
Energy consumption
energy efficiency
Forecasting
Long short term memory
management
Multilayer perceptrons
Optimization
Predictive models
Recurrent neural networks
Smart home
Smart homes
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Each country is looking for effective ways of reducing energy consumption in residential sectors. The process needs close monitoring to comprehend energy usage trends. It is, therefore, important to predict energy consumption in households that will help in formulating efficient ways of optimization and adjustment of energy utilization. The major motivation of this paper is to provide deep learning-based advanced models that would enable the forecasting of household energy consumption with regard to different weather factors. In pursuit of this, we set out to develop a collection of deep learning models Multilayer Perceptron (MLP) and Recurrent Neural Networks (RNN) to evaluate the respective strengths of both the above-named models. The goal is to achieve highly accurate predictions. We evaluated the performance of the MLP and RNN prediction models based on their RMSE scores; the proposed models achieve a lower value of RMSE. Both models correctly recreated the consumption curves with exceptional accuracy.
DOI:10.1109/IC2EM63689.2025.11100757