Evaluation of GMM Clustering Augmentation and Attention Mechanism Integration in LSTM for Electricity Consumption Forecasting

Electrical energy serves as a fundamental pillar of the modern era. Consequently, precise methods are required to forecast fluctuating electricity consumption patterns. This study aims to evaluate the implementation of Gaussian Mixture Model (GMM) clustering and the Attention mechanism within a Long Short-Term Memory (LSTM) framework for electricity consumption prediction. GMM clustering is utilized for feature engineering to uncover contextual patterns within the electricity data. Furthermore, the Attention mechanism, specifically Additive Attention, is employed to measure the relevance of input data. This results in more accurate predictions and reduces the computational burden of processing irrelevant temporal information. The case study utilizes the electricity consumption dataset from Tetuan City, Morocco. The experimental design involves a gradual addition and combination of architectural components such as time features, stacked layers, GMM clustering, and the Attention mechanism. Based on the conducted experiments, the LSTM-Attention model achieved the best MAE and MAPE values compared to other models, recording 428.1213 and 0.68 respectively. Meanwhile, the optimal RMSE value was yielded by the Stacked LSTM-Attention model at 607.0047. Nevertheless, the LSTM-GMM-Attention model also demonstrated competitive performance relative to these two models, with RMSE, MAE, and MAPE values of 610.0044, 437.1562, and 0.70 respectively. This indicates that the integration of architectural components significantly influences the improvement of electricity consumption prediction accuracy. The Attention mechanism effectively filters the most relevant temporal information, thereby enhancing model precision in electricity forecasting. Concurrently, the inclusion of GMM contributes positively to understanding data distribution patterns, although the improvement is not as substantial as the effect produced by the Attention mechanism.

• Publication year

  2025

• Venue

  2025 8th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI)

• Publication date

  2025-12-11

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/ISRITI68345.2025.11393110
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Evaluating the Integration of Fuzzy and Non-Fuzzy Clustering Approaches into LSTM for the Power Consumption Forecasting Utilizing the Case Study Dataset of Tetuan City

  2025cited by this paper
• A hybrid time series forecasting approach integrating fuzzy clustering and machine learning for enhanced power consumption prediction

  2025cited by this paper
• Short-Term Electricity-Load Forecasting by Deep Learning: A Comprehensive Survey

  2024cited by this paper
• Electricity demand forecasting with hybrid classical statistical and machine learning algorithms: Case study of Ukraine

  2024cited by this paper
• Time Series Forecasting using LSTM and ARIMA

  2023cited by this paper
• Perbandingan Prediksi Penggunaan Listrik dengan Menggunakan Metode Long Short Term Memory (LSTM) dan Recurrent Neural Network (RNN)

  2022cited by this paper
• Application of long short-term memory (LSTM) neural network based on deep learning for electricity energy consumption forecasting

  2022cited by this paper
• GMM-LSTM: a component driven resource utilization prediction model leveraging LSTM and gaussian mixture model

  2022cited by this paper
• Forecasting the dynamics of cumulative COVID-19 cases (confirmed, recovered and deaths) for top-16 countries using statistical machine learning models: Auto-Regressive Integrated Moving Average (ARIMA) and Seasonal Auto-Regressive Integrated Moving Average (SARIMA)

  2021cited by this paper
• Quantitative estimation of the impact of climate change on residential electricity demand for the city of Greater Accra, Ghana

  2021cited by this paper
• Global

  2021cited by this paper
• Global Energy Review 2020

  2020cited by this paper
• Electricity consumption prediction based on LSTM with attention mechanism

  2020cited by this paper
• Prediksi Harga Beras Menggunakan Metode Multilayer Perceptron (MLP) dan Long Short Term Memory (LSTM)

  2020cited by this paper
• Improving Electric Energy Consumption Prediction Using CNN and Bi-LSTM

  2019cited by this paper
• Comparison of Machine Learning Algorithms for the Power Consumption Prediction : - Case Study of Tetouan city –

  2018cited by this paper
• Neural Machine Translation by Jointly Learning to Align and Translate

  2014cited by this paper

• No citing papers are available for this paper.