Comparative performance analysis of LSTM and classical machine learning models for IoT-based short-term energy consumption forecasting

Öz

Accurate energy consumption forecasting is essential for sustainable grid management and efficient energy distribution. This study conducts a comparative performance analysis of deep learning and classical machine learning algorithms for short-term energy consumption forecasting using Internet of Things (IoT)-based datasets. Real-time data from smart meters are preprocessed through noise filtering, normalization, and feature selection before being modeled using Long Short-Term Memory (LSTM), Random Forest (RF), Support Vector Regression (SVR), and Extreme Gradient Boosting (XGBoost). Model performance is evaluated using Mean Squared Error (MSE), Mean Absolute Error (MAE), and R² metrics. Results reveal that the LSTM model achieves the highest prediction accuracy due to its superior capability in capturing nonlinear and temporal dependencies in IoT data. The findings highlight the potential of integrating IoT and deep learning to improve real-time decision-making, demand response, and sustainable energy planning in smart grids. Among all tested models, LSTM achieved an R² of 0.983 and MSE of 0.0018, demonstrating its suitability for real-time IoT-based forecasting in smart grids.

Anahtar Kelimeler

• IoT-based energy forecasting
• LSTM
• Machine learning
• XGBoost
• Smart grid
• Short-term prediction

Kaynakça

1. [1] Hong T, Pinson P, Wang Y, Weron R, Yang D, Zareipour H. Energy forecasting: a review and outlook. IEEE Open Access Journal of Power and Energy 2020; 7: 376-388.
2. [2] Wang K, Wang Y, Sun Y, Guo S, Wu J. Big data analytics for smart grids: a survey. IEEE Transactions on Industrial Informatics 2019; 15(2): 1179-1191.
3. [3] Amasyali K, El-Gohary NM. A review of data-driven building energy consumption prediction studies. Renewable and Sustainable Energy Reviews 2018; 81: 1192-1205.
4. [4] Hyndman RJ, Athanasopoulos G. Forecasting: principles and practice. 2nd ed. OTexts, Melbourne, Australia, 2018.
5. [5] Makridakis S, Spiliotis E, Assimakopoulos V. The M4 competition: 100,000 time series and 61 forecasting methods. International Journal of Forecasting 2020; 36(1): 54-74.
6. [6] Ahmad T, Mourshed M, Rezgui Y. Trees vs. neurons: comparison between random forest and ANN for high-resolution building energy load forecasting. Energy and Buildings 2018; 147: 77-89.
7. [7] Chen T, Guestrin C. XGBoost: a scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, US, 2016.
8. [8] Hochreiter S, Schmidhuber J. Long short-term memory. Neural Computation 1997; 9(8): 1735-1780.

9. [9] Bouktif S, Fiaz A, Ouni A, Serhani MA. Optimal deep learning LSTM model for electric load forecasting using feature selection and genetic algorithm. Energies 2018; 11(7): 1636.
10. [10] Kong W, Dong ZY, Jia Y, Hill DJ, Xu Y, Zhang Y. Short-term residential load forecasting based on LSTM recurrent neural network. IEEE Transactions on Smart Grid 2017; 10(1): 841-851.
11. [11] da Silva IN, Ribeiro MHDM, Mariani VC, Coelho LS. Forecasting short-term electricity demand using hybrid LSTM-ANN approaches. Energy 2019; 169: 667-678.
12. [12] Bandara K, Bergmeir C, Hewamalage H. LSTM-MSNet: multi-horizon forecasting with long short-term memory networks. Pattern Recognition Letters 2020; 136: 72-79.
13. [13] UCI Machine Learning Repository. Individual household electric power consumption data set. 2006. [Online] Available from: https://archive.ics.uci.edu/dataset/235/individual+household+electric+power+consumption (Accessed: 15/05/2026).
14. [14] Kelly J, Knottenbelt W. The UK-DALE dataset, domestic appliance-level electricity demand and whole-house demand. Scientific Data 2015; 2: 150007.
15. [15] Murray D, Stankovic L, Stankovic V, Lulic S, Sluganovic I, Gray S. A dataset of household electric power demand in the UK. Scientific Data 2017; 4: 170122.
16. [16] Open Power System Data. Time series and weather data for Europe. 2017. [Online] Available from: https://open-power-system-data.org/ (Accessed: 15/05/2026).
17. [17] Niu D, Shi H, Wang M, Chan FTS, Li D. Electricity load forecasting using a deep learning network in smart grid. Journal of Intelligent & Fuzzy Systems 2020; 38(6): 7611-7624.
18. [18] Zhang GP. Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing 2003; 50: 159-175.
19. [19] Marino DL, Amarasinghe K, Manic M. Building energy load forecasting using deep neural networks. Energy Procedia 2016; 105: 2643-2648.
20. [20] Oreshkin BN, Carpov D, Chapados N, Bengio Y. N-BEATS: neural basis expansion analysis for interpretable time series forecasting. International Journal of Forecasting 2020; 36(3): 1091-1102.
21. [21] Lim B, Arık SÖ, Loeff N, Pfister T. Temporal fusion transformers for interpretable multi-horizon time series forecasting. Nature Machine Intelligence 2021; 3: 105-113.
22. [22] Ribeiro MHDM, Coelho LS, da Silva IN. Short-term forecasting in power systems using LSTM networks and wavelet decomposition. Electric Power Systems Research 2020; 179: 106058.
23. [23] Ahmad T, Chen H, Guo Y, Wang J, Haleem A. A comprehensive review on energy forecasting: modeling approaches and future research directions. Energies 2020; 13(15): 3863.
24. [24] Kuo PH, Huang CJ. A high-precision ANN model for short-term energy load forecasting. Energies 2018; 11(1): 213.
25. [25] Taylor SJ, Letham B. Forecasting at scale. The American Statistician 2018; 72(1): 37-45.
26. [26] Lago J, De Ridder F, De Schutter B. Forecasting spot electricity prices: deep learning approaches and empirical comparison. Applied Energy 2018; 221: 386-405.
27. [27] Wang H, Lei Z, Zhang X, Zhou B, Peng J. A review of deep learning for renewable energy forecasting. Energy Conversion and Management 2019; 198: 111799.
28. [28] Dada EG, Oyewola DO, Yakubu JH. Power consumption prediction in urban areas using machine learning as a strategy towards smart cities. Arid Zone Journal of Basic and Applied Research 2022; 1(1): 11-24.
29. [29] Bouktif S, Ouni A, Serhani MA. Forecasting household electricity consumption using hybrid deep learning models. Applied Energy 2020; 263: 114682.
30. [30] Maarif MR, Saleh AR, Habibi M, Fitriyani NL, Syafrudin M. Energy usage forecasting model based on LSTM and explainable artificial intelligence. Information 2023; 14(5): 265.
31. [31] Natarajan Y, et al. Enhancing building energy efficiency with IoT-driven hybrid deep learning models for accurate energy consumption prediction. Sustainability 2024; 16(5): 1925.
32. [32] Ryu S, Noh J, Kim H. Deep neural network-based demand side short-term load forecasting. Energies 2017; 10(1): 3.
33. [33] Deb C, Zhang F, Yang J, Lee SE, Shah KW. A review on time series forecasting techniques for building energy consumption. Renewable and Sustainable Energy Reviews 2017; 74: 902-924.
34. [34] Wang H, et al. Energy consumption prediction using modified deep CNN-BiLSTM model. Heliyon 2021; 7(4): e06637.
35. [35] Maarif MR, et al. Predicting short-term energy usage in a smart home using hybrid deep learning models. Frontiers in Energy Research 2024; 12: 1323357.
36. [36] Shi H, Xu M, Li R. Deep learning for household load forecasting-a novel pooling deep RNN. IEEE Transactions on Smart Grid 2017; 9(5): 5271-5280.
37. [37] Wang H, Lei Z, et al. A review of deep learning for renewable energy forecasting. Energy Conversion and Management 2019; 198: 111799.
38. [38] Ahmad T, Chen H, Guo Y, Wang J, Haleem A. A comprehensive review on energy forecasting: modeling approaches and future research directions. Energies 2020; 13(15): 3863.
39. [39] Herzen J, Lässig J, Norel R, et al. Darts: user-friendly forecasting with Python. Journal of Machine Learning Research 2022; 23(124): 1-6.
40. [40] Hong T, Pinson P, Wang Y. Energy forecasting in the era of big data analytics and AI. International Journal of Energy Research 2021; 45(12): 17760-17782.