cs.LGApr 14, 2026

Beyond Weather Correlation: A Comparative Study of Static and Temporal Neural Architectures for Fine-Grained Residential Energy Consumption Forecasting in Melbourne, Australia

Authors: Prasad Nimantha Madusanka Ukwatta Hewage, Hao Wu

Organizations: School of Computing and Information Technology Victoria University, Sydney, NSW, Australia

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Accurate short-term residential energy consumption forecasting at sub-hourly resolution is critical for smart grid management, demand response programmes, and renewable energy integration. While weather variables are widely acknowledged as key drivers of residential electricity demand, the relative merit of incorporating temporal autocorrelation - the sequential memory of past consumption; over static meteorological features alone remains underexplored at fine-grained (5-minute) temporal resolution for Australian households. This paper presents a rigorous empirical comparison of a Multilayer Perceptron (MLP) and a Long Short-Term Memory (LSTM) recurrent network applied to two real-world Melbourne households: House 3 (a standard grid-connected dwelling) and House 4 (a rooftop solar photovoltaic-integrated household). Both models are trained on 14 months of 5-minute interval smart meter data (March 2023-April 2024) merged with official Bureau of Meteorology (BOM) daily weather observations, yielding over 117,000 samples per household. The LSTM, operating on 24-step (2-hour) sliding consumption windows, achieves coefficients of determination of R^2 = 0.883 (House 3) and R^2 = 0.865 (House 4), compared to R^2 = -0.055 and R^2 = 0.410 for the corresponding weather-driven MLPs - differences of 93.8 and 45.5 percentage points. These results establish that temporal autocorrelation in the consumption sequence dominates meteorological information for short-term forecasting at 5-minute granularity. Additionally, we demonstrate an asymmetry introduced by solar generation: for the PV-integrated household, the MLP achieves R^2 = 0.410, revealing implicit solar forecasting from weather-time correlations. A persistence baseline analysis and seasonal stratification contextualise model performance. We propose a hybrid weather-augmented LSTM and federated learning extensions as directions for future work.

Figures & tables

VariableAvailableSelectedRationale
Minimum temperature ( ∘ C)✓×\timesCorrelated with 9 am temp
Maximum temperature ( ∘ C)✓✓Key cooling/heating driver
Rainfall (mm)✓✓Proxy for cloud cover
Evaporation (mm)✓×\times>> 40% missing values
Sunshine (hours)✓×\times100% missing — dropped
9 am Temperature ( ∘ C)✓✓Morning demand indicator
Table 1: BOM Variables: Available vs. Selected
AttributeHouse 3House 4
Temporal resolution5 minutes5 minutes
Date range2023-03-01 to 2024-04-172023-03-01 to 2024-04-17
Total records117,513119,100
Training records (80%)94,01095,280
Test records (20%)23,50323,820
Target variableGrid consumption (W)Total consumption (W)
Table 2: Summary Statistics of Final Merged Datasets
Figure 1: Neural network architectures: (a) Multilayer Perceptron mapping static weather features to consumption; (b) LSTM processing a 2-hour sliding window of consumption history.
 House 3 (Grid Only)House 4 (Solar PV)
MetricTrainTestTrainTest
RMSE (W)752.60971.14830.70673.59
MAE (W)524.72661.39492.71415.51
R2R^{2}0.020−- 0.0550.3040.410
Epochs (early stop)52≈{\approx} 40
Table 3: MLP Evaluation Results
MetricHouse 3House 4
Test Loss (MSE, scaled)0.0012—
Test MAE (scaled)0.0183—
Test R2R^{2}0.8830.865
Training windows181,497181,776
Test windows45,37545,504
Sequence length24 steps (2 hours)24 steps (2 hours)
Table 4: LSTM Evaluation Results
ModelHouse 3 R2R^{2}House 4 R2R^{2}Features Used
Naïve Persistence+0.878+0.878+0.851+0.851Last 5-min value only
Seasonal Naïve−0.401-0.401+0.213+0.213Same slot, prior day
MLP−0.055-0.055+0.410+0.410Weather ×\times 6 + Time
LSTM+0.883\mathbf{+0.883}+0.865\mathbf{+0.865}24-step sequence
Table 5: Complete Model Comparison — Test Set R2R^{2}
Figure 2: Model performance comparison ( R2R^{2} ) across all architectures and both households. The LSTM achieves an improvement of 93.3 percentage points over the MLP on House 3. Note: Positive R2R^{2} values indicate explanatory power above the mean baseline; negative values indicate performance below the trivial mean predictor.
Figure 3: House 3 test set: actual consumption vs. model predictions over a representative 3-day window. (a) Naïve persistence; (b) MLP weather-driven predictions; (c) LSTM sequence predictions. The LSTM (c) tracks actual peaks and troughs closely, whereas the MLP (b) produces a near-constant prediction.
Figure 4: House 4 (Solar PV): grid draw, solar generation, and total consumption over a 5-day test window. Solar generation follows a deterministic daily arc correlated with weather, partially explaining the MLP’s improved performance ( R2=0.410R^{2}=0.410 ) compared to House 3.
Figure 5: House 3 seasonal analysis: (a) consumption distribution by season; (b) median diurnal profiles by season. Summer shows greater spread (air conditioning variability); winter shows elevated baseline (heating). Morning/evening peaks are present in all seasons.
Figure 6: Pearson correlation heatmaps: (a) House 3 weather features vs. grid consumption; (b) House 4 weather features vs. total consumption. House 4 shows stronger weather–load correlations due to the solar generation component.
VenueTypeIFStage
arXiv cs.LGPreprint—Immediate
IEEE PowerTech / ISGTConference—3 months
Applied EnergyJournal11.26 months
Energy and BuildingsJournal6.76 months
IEEE Trans. Smart GridJournal8.969 months
Renewable & Sust. Energy ReviewsJournal16.112 months
Table 6: Target Publication Venues
Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

FileMonthRecords
202303.csvMarch 202331
202304.csvApril 202330
202305.csvMay 202331
202306.csvJune 202330
202307.csvJuly 202331
202308.csvAugust 202331
Appendix
Table 7: Bureau of Meteorology Monthly Files
More in this topic

Electricity Load Forecasting

Electricity load forecasting predicts future power demand to support grid operation, planning, and energy management.

See all 46 papers

Explore similar work

CardsList
Aug 19, 2026

An Empirical Benchmark of Deep Time-Series Models for Smart Meter Energy Forecasting

Accurate forecasting of energy consumption is important for the efficient operation of power systems, with direct implications for operational costs, energy management, and system maintenance. Due to the availability of extensive high-resolution consumption data from smart meters, data-driven methods have been used for short-term and long-term forecasting. However, their comparative performance on real-world smart meter data is still not well studied. In this paper, we present an empirical benchmark of nine modern deep learning models for time-series forecasting, including linear, MLP-based, convolutional, and Transformer architectures. We evaluate these models on two publicly available smart meter datasets. Our analysis focuses on three factors that strongly affect forecasting performance: the length of historical input, the prediction horizon, and the choice of model architecture. We show that extending the historical context improves accuracy, but only up to a saturation point, after which additional input provides limited benefit. In contrast, accuracy decreases as the prediction horizon increases. We also investigate the trade-off between prediction accuracy and computational complexity, and assess the statistical significance and practical magnitude of performance differences across models. Our results show that deep learning models consistently outperform classical baselines, while lightweight architectures achieve relatively similar performance at significantly lower computational cost. Additionally, architectural differences only become meaningful at longer forecasting horizons and on more heterogeneous datasets. Finally, a subgroup analysis across geodemographic and household categories shows that model choice has limited impact for most population segments.

Behnaz Kavoosighafi, Maria Eidenskog, Wiktoria Glad +1

Linköping University, Department of Science and Technology, Norrköping, 60174, Östergötland, Sweden · Linköping University, Department of Thematic Studies, Linköping, 58183, Östergötland, Sweden

Time Series ForecastingForecasting BenchmarksElectricity Load Forecasting
AskPDF
Oct 8, 2026

Machine Learning and Time-Series Foundation Models for Day-Ahead Electricity Forecasting in a University Building

Day-ahead electricity forecasts can help building operators schedule heating, ventilation, and other services around anticipated demand. Choosing a forecasting model also requires balancing accuracy with its forecast-time inputs, local training, and computational requirements. We benchmark statistical, machine-learning, recurrent neural-network, and pretrained TimesFM models for a university building. We use chronological validation and evaluate all configurations on the same 57 test days, comparing accuracy across workdays, weekends, and holidays alongside local fitting and forecasting time. Weather covariates use retrospective observations. TimesFM 3.0 with covariates achieves the lowest overall errors, reducing mean absolute error by approximately 43% relative to weekly seasonal-naive forecasting. XGBoost remains competitive, although paired daily comparisons do not establish a clear mean-absolute-error advantage for TimesFM. TimesFM also performs strongly using only past electricity consumption. Adding covariates improves its holiday forecasts but slightly worsens workday and weekend forecasts, leaving overall mean absolute error almost unchanged. TimesFM requires no building-specific weight training, while XGBoost offers rapid local fitting and inference. The evaluated LSTM takes longer to fit without outperforming either approach. An exploratory historical evaluation reveals seasonal variation, while a separate weather-source comparison produces only small changes in the selected TimesFM configuration's accuracy. These results support pretrained TimesFM as a strong initial forecasting baseline for this building, with XGBoost as a competitive locally fitted alternative. Covariates are particularly useful when improved holiday forecasts justify the additional input preparation.

Amey Vasulkar, Samuel Kernan Freire, Arie-Willem de Leeuw +1

Electricity Load ForecastingForecasting BenchmarksTime-Series Foundation Models
AskPDF
Apr 26, 2026

Machine Learning and Deep Learning Models for Short Term Electricity Price Forecasting in Australia's National Electricity Market

Short term electricity price forecast is essential in competitive power markets, yet electricity price series exhibit high volatility, irregularity, and non-stationarity. This phenomenon is pronounced in the South Australian region of the National Electricity Market, where high renewable penetration drives price volatility and frequent negative price intervals, while structural changes such as the transition to five-minute settlement further complicate forecast. To address these challenges, this study develops a unified benchmark framework. Under identical data preprocessing, feature engineering with lag features, rolling statistics, cyclic temporal encodings, and so on, and an 85% to 15% chronological train test split, six algorithms are systematically compared, including AWMLSTM, CatBoost, GBRT, LSTM, LightGBM, and SVR. The results show that for price prediction, tree-based models, especially GBRT with an R squared value of 0.88, generally outperform LSTM and SVR. However, all models achieve a mean absolute percentage error above 90%, and more than 65% of GBRT predictions have relative errors above 10%, which highlights the inherent difficulty of price forecast. For demand prediction, all models perform substantially better than in price prediction. AWMLSTM and GBRT achieve an R2 value of 0.96 with mean absolute percentage error below 32%, and GBRT has 74.37% of samples within 5% error, while LSTM and SVR perform less accurately in both tasks. Future improvements should focus on hybrid models such as tree plus transformers, data augmentation for extreme events, and error correction to better capture price spikes.

Wei Lu, Jay Wang, Dingli Duan +3

School of Engineering, Computer & Mathematical Sciences, Auckland University of Technology (AUT), Auckland 1010, New Zealand · Department of Thermal Energy and Power Engineering, Yantai University, Yantai, China · Department of Building Environment and Equipment, School of Civil Engineering, Hefei University of Technology, Hefei, China +1

Gradient-Boosted Decision TreesElectricity Price ForecastingTime Series ForecastingForecasting Benchmarks
AskPDF