Pith. sign in

REVIEW 4 major objections 5 minor 74 references

Data-Driven Heat Pump Management: Combining Machine Learning with Anomaly Detection for Residential Hot Water Systems

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read On six real Swedish households, a LightGBM forecaster predicts mid-tank hot-water temperature more accurately than LSTM variants, and an isolation forest turns those predictions into a household-specific shower-event calendar, supporting…

desk verdict A useful real-world forecasting benchmark undercut by unsupported anomaly-detection metrics and an unresolved data-split contradiction. read the letter →

arxiv 2506.15719 v1 pith:YNCPHIAK submitted 2025-06-03 cs.LG

Manal Rahal , Bestoun S. Ahmed , Roger Renstrom , Robert Stener , Albrecht Wurtz This is my paper
classification cs.LG
keywords MachinelearningHeatpumpHotwaterdemandforecastingLightGBMLSTMIsolationforestAnomalydetectionDemand-responsivecontrol
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper aims to establish that a heat pump can manage residential hot water production by learning each household's consumption pattern instead of reacting to fixed temperature thresholds. The proposed pipeline predicts the temperature at the middle of the water tank with machine learning, then uses isolation-forest anomaly detection to mark the large temperature drops that correspond to shower events, and finally builds a per-household calendar of expected hot water use. The results on six real Swedish households are the point: LightGBM outperforms LSTM-based models in root-mean-square error by up to 9.37% with $R^2$ from 0.748 to 0.983, and the isolation forest finds shower events with F1-score 0.87 and a false-alarm rate of 5.2%. If these numbers hold, the same two-stage scheme could steer heat pump start and stop commands toward predicted demand, reducing wasted production and keeping users from running out of hot water.

What carries the argument

The load-bearing object is the mid-tank temperature $t_{\text{mid}}$, which drops sharply when hot water is drawn and therefore encodes both the household's demand and the start/stop needs of the heat pump. The models are trained on lagged temperature features $t_{\text{mid}}(10)$, $t_{\text{mid}}(20)$, $t_{\text{mid}}(30)$, $t_{\text{mid}}(90)$, the corresponding $t_{\text{top}}$ lags, current $t_{\text{top}}$, and the week number. LightGBM, a gradient-boosting tree method using gradient-based one-side sampling and exclusive feature bundling, supplies the forecast; isolation forest, which scores each point by the path length needed to isolate it in random decision trees, supplies the anomaly detection that identifies shower events. The paper's mechanism is the combination: forecast first, then isolate unusually sharp temperature drops, then aggregate the detected events into a demand calendar that decides when the heat pump should start producing hot water.

What would settle it

Inspect the preprocessing pipeline or rerun it exactly, and test for leakage by checking whether any row in the test period contributed to the lagged values used for training. Concretely, retrain LightGBM on data split strictly by timestamp with no shuffling and recompute RMSE and $R^2$; if the numbers match the paper's, the split was clean, and if they are materially worse, the published scores are inflated by future information.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a lightweight gradient-boosting forecaster, LightGBM, predicts the mid-tank temperature of a residential hot water tank more accurately than recurrent models (LSTM and bidirectional LSTM with self-attention), and that an unsupervised isolation forest can then turn those predictions into a reliable household-specific calendar of shower events. The claim is that this two-stage design makes demand-responsive heat pump control practical without per-household thresholds: the model consumes ten lagged sensor values for the top and middle tank temperatures plus the current top temperature and week number, predicts future mid-tank temperature, and the iForest flags the sharp drops that indicate hot water use. Across six installations the RMSE improvement over LSTM reaches 9.37%, and the shower-event detector reaches F1 0.87 with a false-alarm rate of 5.2%.

Load-bearing premise

The reported accuracy depends on the 85/15 split being chronological and leakage-free; the data-processing diagram's label 'Shuffling and random split of data' conflicts with the date-based train/test ranges in the experimental setup, and if shuffling preceded the split, the lag features would expose near-future values during training.

Editorial extensions

If this is right

  • A heat pump controller that knows the household's predicted demand can start hot water production before expected shower times rather than only after temperature falls below a threshold.
  • The same trained forecaster can be retuned per household, since the paper finds household-specific LightGBM tuning is needed while the anomaly detector carries over without per-household thresholds.
  • The two-stage pipeline could be applied to other appliances or time series where a physical variable, such as temperature, pressure, or flow, drops sharply during usage events.
  • Because LightGBM trains in about a third of the time of the LSTM variants, the approach is more feasible for resource-constrained edge controllers.
  • The shower-event calendar gives a concrete weekly operating schedule from the data, not from fixed assumptions about when people use hot water.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: because the paper never measures electricity consumption or energy savings from the generated calendar, the practical payoff of the approach would need a randomized field trial that compares threshold-based and forecast-based control under matched weather and occupancy.
  • Inference: the split inconsistency should be resolved before the numeric claims are taken at face value; a strictly chronological rerun is the first check a deploying engineer would run.
  • Inference: the forecast-then-isolate pattern is portable to other usage-correlated physical signals, such as flow, pressure, or appliance current draw, but the weekly calendar construction assumes fairly regular household routines and would need a separate treatment for irregular consumers.
  • Inference: the paper's own admission that no ablation study was run leaves open which features, especially the ten-minute lag versus the full lag set, carry the forecast skill; a feature ablation would be an inexpensive next experiment.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a data-driven pipeline for residential heat-pump hot-water management: three machine-learning models (LightGBM, LSTM, and a BiLSTM with self-attention, called attLSTM) are trained to forecast the mid-tank temperature tmid, and an isolation-forest (iForest) module is applied to the forecast to detect shower events, from which a household-specific hot-water demand calendar is derived. The evaluation is conducted on six real Swedish households. The paper reports that LightGBM outperforms the LSTM variants with RMSE improvements up to 9.37% and R2 values between 0.748 and 0.983, and that the iForest anomaly detector achieves an F1-score of 0.87 with a false alarm rate of 5.2%. The central claim is that this composite approach can steer heat-pump hot-water production more responsively than conventional threshold-based control.

Significance. If the reported results are valid, the paper would provide a useful application-level contribution: real sensor data from six households, a comparison of three model families with explicit hyperparameter search spaces, and a concrete anomaly-detection step aimed at producing shower-event calendars. The paper also reports computational training times and documents the data-preprocessing choices, which is helpful for reproducibility. However, the two headline results are not yet supported. The anomaly-detection metrics appear only in the abstract and introduction, with no labeled evaluation in the experimental sections. The forecasting evaluation is threatened by an unresolved contradiction about whether the data were shuffled before the chronological split, which would invalidate the reported error metrics. These are load-bearing gaps rather than presentation issues.

major comments (4)
  1. [Figure 2; Section 5.2.1; Section 6.1; Table 2] There is an unresolved contradiction about the data split that directly affects the validity of Table 7. Figure 2 lists 'Shuffling and random split of data' inside the Data Processing block, while Section 5.2.1 and Section 6.1 describe an 85/15 split of the data into training and testing, and Table 2 gives contiguous chronological date ranges for the two parts. Because the feature set in Table 3 includes lagged values tmid(10), tmid(20), tmid(30), and tmid(90) and the target is a future value of tmid, a random shuffle before the split would place test rows within minutes of training rows. Given the strong autocorrelation of tank temperature, the model could then effectively copy or interpolate near-identical sensor readings rather than forecast, inflating the reported R2, RMSE, and MAPE values. The authors must clarify whether rows were shuffled before or after the split, and if shuffling was applied, the evaluation must be rerun on a strictly chronological split.
  2. [Abstract and Section 1 vs. Section 6.5] The claimed anomaly-detection performance, F1-score of 0.87 and false alarm rate of 5.2%, is reported in the abstract and introduction but is never derived in the methods or results sections. Section 6.5 applies iForest to the forecast window for household 4 and shows detected shower events in Figure 8, but no ground-truth labels, confusion matrix, precision/recall computation, or evaluation equation are provided. The abstract's quantitative claim is therefore unsupported. The authors should either add an explicit anomaly-detection evaluation protocol with labeled events or remove these numbers from the headline claims.
  3. [Section 5.1.5 and Section 7] The false alarm rate of 5.2% appears to be essentially determined by the hand-set iForest contamination rate. Section 5.1.5 states that the contamination rate is preset to 0.05 for all households, and Section 7 repeats that the contamination rate was chosen after experimentation but without reporting those experiments. With no labeled anomalies, the false alarm rate cannot be computed independently, and the F1-score is not defined. The authors should state how false alarms and F1 were computed, and how the choice of contamination rate affects these metrics.
  4. [Section 6.3 and Table 6] Table 6 does not report the final hyperparameter values used for LightGBM; it reproduces the search ranges (e.g., 'Max Depth 5, 10, 30', 'Learning Rate 0.001, 0.01, 0.1') rather than the selected values for each household. This makes the headline model comparison irreproducible. The final selected hyperparameters for LightGBM should be reported for each household, as is done for the LSTM models.
minor comments (5)
  1. [Section 5.2.2] The feature-selection step is described as using 'partial OLS regression', but the method is not defined or cited; please clarify what 'partial' means here so that the feature-selection procedure can be reproduced.
  2. [Section 6.5] The text refers to 'Figure 6.5' before Figure 8; the figure numbering should be corrected.
  3. [Section 8] The section states that limitations are categorized into two main areas, but then lists three categories: model-specific limitations, data quality dependencies, and dataset considerations.
  4. [Throughout] The notation is inconsistent: 'R2' should be written as R^2 or R², and the attention-based model is variously called 'attLSTM', 'BiLSTM with attention', and 'Bi-LSTM with Attention'.
  5. [Section 6.2] Equation (11) contains an odd LaTeX rendering ('/radicaltp/radicalvertex/radicalvertex√') that should be replaced with a proper square-root symbol.

Circularity Check

1 steps flagged · score 6.0 of 10

Anomaly-detection metrics reduce to the hand-set iForest contamination rate; the forecast comparison is not circular, though its split is leakage-ambiguous.

  1. fitted input called prediction [Section 5.1.5 (contamination), Section 8 (Limitations, tuning), Abstract (F1/FAR), Section 7 (unsupervised choice)]
    "'The contamination rate chosen for all the use cases in this paper is 0.05.' ... 'we selected a contamination rate of 0.05 after rounds of experimentation with multiple different rates.' ... 'For anomaly detection, our iForest implementation achieved an F1-score of 0.87 with a false alarm rate of only 5.2%.' ... 'data labeling for every household is typically challenging and expensive to perform... iForest as an unsupervised anomaly detection method is preferred.'"

    The 0.05 contamination parameter tells iForest what fraction of the training data is anomalous; in iForest the contamination value sets the decision threshold, so the proportion flagged is 5% by construction. The paper's 'false alarm rate of only 5.2%' is therefore the chosen contamination prior (plus the 30-minute post-processing filter), not an independently measured error rate. Since Section 7 justifies iForest as unsupervised because labeling is 'challenging and expensive,' no external ground-truth label set for shower events is described; the F1-score of 0.87 cannot be verified against independent labels and is consistent with being computed from the model's own anomaly assignments.

full rationale

The forecasting contribution is not circular: inputs are past and current sensor values (tmid(90), tmid(30), tmid(20), tmid(10), ttop(90), ttop(30), ttop(20), ttop(10), ttop, Week) and the target is a future tmid value; no equation defines the prediction as one of its inputs. The unresolved 'Shuffling and random split of data' statement in Figure 2 versus the chronological dates and 85/15 split in Section 5.2.1/Table 2 is a potential data-leakage threat to Table 7, but leakage is a correctness problem, not circularity. The circularity is localized to the anomaly-detection headline: Section 5.1.5 fixes the iForest contamination rate to 0.05 for all households, Section 8 says this rate was selected 'after rounds of experimentation,' and the abstract reports a false alarm rate of only 5.2%. Because contamination directly sets the fraction of data iForest labels anomalous, the reported false alarm rate is essentially the tuned input restated as a measured outcome. Section 7 also justifies iForest as unsupervised because labeling is 'challenging and expensive,' and no external ground-truth labels or F1 computation protocol are described for shower events, so the 0.87 F1-score is not anchored to an independent label set. The only self-citation (ref. 11, co-authored by B.S. Ahmed) appears in related-work context and is not load-bearing. Overall: one of the two headline results is partially self-referential; the forecasting comparison retains independent content. Score 6.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central evaluation rests on a handful of hand-set parameters and domain assumptions. The most important is the iForest contamination rate, set to 0.05, which directly produces the reported false alarm rate. The paper also assumes the data split is free of temporal leakage, that forward-fill preserves the signal, and that tmid is a valid proxy for hot water demand. No new physical or mathematical entities are introduced.

free parameters (3)
  • iForest contamination rate = 0.05
    Set manually for all six households (Section 5.1.5); the reported false alarm rate of 5.2% is essentially this parameter, making the anomaly-detection evaluation circular.
  • Post-processing window for merging shower events = 30 minutes
    Ad hoc filter to avoid consecutive detection of a single shower event (Section 5.1.5); not justified by data or independent validation.
  • Lag set for forecasting features = {10, 20, 30, 90} minutes
    Selected via OLS p-values (Section 5.2.2); no ablation and no justification that these lags are optimal for the operational horizon.
assumptions (4)
  • domain assumption Sensor data recorded only on change are accurately reconstructed by forward-fill.
    Preprocessing (Section 5.2.1) fills missing timestamps by forward fill; if actual temperature changes occurred during the gaps, the filled values distort the time series and the learned lags.
  • domain assumption tmid is a sufficient proxy for household hot water demand.
    The whole pipeline predicts tmid and treats its drops as shower events; no direct flow or consumption measurement is used, so any consumption not reflected in tmid (e.g., small draws) is missed.
  • domain assumption The 85/15 split is chronological and leakage-free.
    Figure 2 mentions shuffling while Table 2 shows date ranges; the paper never explicitly rules out random splitting, which would leak future information into training.
  • domain assumption iForest anomaly scores correspond to actual shower events without labeled ground truth.
    The paper reports F1-score 0.87 but does not describe how shower events were labeled or how false positives and negatives were counted (Section 6.5 only shows visual examples).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data-Driven Heat Pump Management: Combining Machine Learning with Anomaly Detection for Residential Hot Water Systems." pith.science (2026). https://pith.science/paper/YNCPHIAK

@misc{pith2026250615719,
  author       = {Pith},
  title        = {Pith review of: Data-Driven Heat Pump Management: Combining Machine Learning with Anomaly Detection for Residential Hot Water Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YNCPHIAK}},
  note         = {Machine review of arXiv:2506.15719}
}
abstract

Heat pumps (HPs) have emerged as a cost-effective and clean technology for sustainable energy systems, but their efficiency in producing hot water remains restricted by conventional threshold-based control methods. Although machine learning (ML) has been successfully implemented for various HP applications, optimization of household hot water demand forecasting remains understudied. This paper addresses this problem by introducing a novel approach that combines predictive ML with anomaly detection to create adaptive hot water production strategies based on household-specific consumption patterns. Our key contributions include: (1) a composite approach combining ML and isolation forest (iForest) to forecast household demand for hot water and steer responsive HP operations; (2) multi-step feature selection with advanced time-series analysis to capture complex usage patterns; (3) application and tuning of three ML models: Light Gradient Boosting Machine (LightGBM), Long Short-Term Memory (LSTM), and Bi-directional LSTM with the self-attention mechanism on data from different types of real HP installations; and (4) experimental validation on six real household installations. Our experiments show that the best-performing model LightGBM achieves superior performance, with RMSE improvements of up to 9.37\% compared to LSTM variants with $R^2$ values between 0.748-0.983. For anomaly detection, our iForest implementation achieved an F1-score of 0.87 with a false alarm rate of only 5.2\%, demonstrating strong generalization capabilities across different household types and consumption patterns, making it suitable for real-world HP deployments.

Figures

Figures reproduced from arXiv: 2506.15719 by the authors.

Figure 1
Figure 1. The HP water tank operations comfortable experience for the user. The HPs included in this study start and stop hot water production based on preset thresholds for the average temperature between tmid and top ttop. As a result of the existing system, production may start when there is low or no demand for hot water due to the natural decline of temperature in the tank, such as after midnight. This leads to a major d… view at source ↗
Figure 2
Figure 2. Framework for forecasting and adapting the production of hot water production to the household demand. replaces the conventional threshold-based method that steers the hot water production in an HP. This is achieved by predicting tmid for a specific time window using ML, then applying an anomaly detection method to build the household demand calendar and steer the production of hot water in an HP. The integration of… view at source ↗
Figure 3
Figure 3. Overview of the household data preprocessing steps. of the influential input variables. In [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Distribution of measurements of sensors per household current ttop is also included to provide the model with the latest state of the water tank temperature profile. As for the week predictor, it is included as an integer to account for seasonal variations that may inf…
Figure 5
Figure 5. Figure 5: Overall MAPE and RMSE for the six households. 0 10 20 30 40 50 Epoch 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 MAE Training Loss Validation Loss (a) Household 4 0 10 20 30 40 50 Epoch 1 2 3 4 MAE Training Loss Validation Loss (b) Household 5 [PITH_FULL_IMAGE:figures/full_fi…
Figure 6
Figure 6. Figure 6: LSTM learning curves for household 4 and 5 certain consumption patterns, where the test suggests no significant difference between attLSTM and LightGBM (p=0.0938). In [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]
Figure 7
Figure 7. Figure 7: AttLSTM learning curves for household 4 and 5 23 [PITH_FULL_IMAGE:figures/full_fig_p023_7.png]
Figure 6
Figure 6. Figure 6: shows the actual and forecasted temperatures [PITH_FULL_IMAGE:figures/full_fig_p024_6.png]
Figure 8
Figure 8. Figure 8: Detection of shower events using iForest. extracted, where weekdays are encoded as integers from 0 (Monday) to 6 (Sunday). In addition, the hour is extracted from the timestamp. This allows us to aggregate the frequency of shower events by weekday and hour. This aggreg…
Figure 9
Figure 9. Figure 9: Hourly probability of shower events across weekdays in household 4. and detect the expected shower events respectively. The selection of iForest in the case of shower event detection enables our approach to dynamically respond to changing behavioral patterns among the …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

74 extracted references · 74 canonical work pages

  1. [1]

    Heat pumps and our low-carbon future: A comprehensive review, Energy Research & Social Science 2021, 71, 101764

    Gaur, A.S.; Fitiwi, D.Z.; Curtis, J. Heat pumps and our low-carbon future: A comprehensive review, Energy Research & Social Science 2021, 71, 101764

    work page 2021

  2. [2]

    Energy flexible heat pumps in industrial energy systems: A review, Energy Reports 2023, 9, 386--394

    Sadjjadi, B.S.; Gerdes, J.N.; Sauer, A. Energy flexible heat pumps in industrial energy systems: A review, Energy Reports 2023, 9, 386--394

    work page 2023

  3. [3]

    Machine Learning and Deep Learning in Energy Systems: A Review, Sustainability 2022, 14 (8)

    Forootan, M.M.; Larki, I.; Zahedi, R.; Ahmadi, A. Machine Learning and Deep Learning in Energy Systems: A Review, Sustainability 2022, 14 (8)

    work page 2022

  4. [4]

    A review on renewable energy and electricity requirement forecasting models for smart grid and buildings, Sustainable Cities and Society 2020, 55, 102052

    Ahmad, T.; Zhang, H.; Yan, B. A review on renewable energy and electricity requirement forecasting models for smart grid and buildings, Sustainable Cities and Society 2020, 55, 102052

    work page 2020

  5. [5]

    Optimizing renewable energy systems through artificial intelligence: Review and future prospects, Energy & Environment 2024, 35 (7), 3833--3879

    Ukoba, K.; Olatunji, K.O.; Adeoye, E.; Jen, T.C.; Madyira, D.M. Optimizing renewable energy systems through artificial intelligence: Review and future prospects, Energy & Environment 2024, 35 (7), 3833--3879

    work page 2024

  6. [6]

    Heating up the global heat pump market, Nature Energy 2022, 7 (10), 901--904

    Rosenow, J.; Gibb, D.; Nowak, T.; Lowes, R. Heating up the global heat pump market, Nature Energy 2022, 7 (10), 901--904

    work page 2022

  7. [7]

    Estimating electric power consumption of in-situ residential heat pump systems: A data-driven approach, Applied Energy 2023, 352, 121971

    Song, Y.; Peskova, M.; Rolando, D.; Zucker, G.; Madani, H. Estimating electric power consumption of in-situ residential heat pump systems: A data-driven approach, Applied Energy 2023, 352, 121971

    work page 2023

  8. [8]

    Artificial intelligence and machine learning approaches to energy demand-side response: A systematic review, Renewable and Sustainable Energy Reviews 2020, 130, 109899

    Antonopoulos, I.; Robu, V.; Couraud, B.; Kirli, D.; Norbu, S.; Kiprakis, A.; Flynn, D.; Elizondo-Gonzalez, S.; Wattam, S. Artificial intelligence and machine learning approaches to energy demand-side response: A systematic review, Renewable and Sustainable Energy Reviews 2020, 130, 109899

    work page 2020

Show all 74 references
  1. [9]

    Data-driven soft sensors targeting heat pump systems, Energy Conversion and Management 2023, 279, 116769

    Song, Y.; Rolando, D.; Avellaneda, J.M.; Zucker, G.; Madani, H. Data-driven soft sensors targeting heat pump systems, Energy Conversion and Management 2023, 279, 116769

    2023

  2. [10]

    Recent advances in the analysis of residential electricity consumption and applications of smart meter data, Applied Energy 2017, 208, 402--427

    Yildiz, B.; Bilbao, J.; Dore, J.; Sproul, A. Recent advances in the analysis of residential electricity consumption and applications of smart meter data, Applied Energy 2017, 208, 402--427

    2017

  3. [11]

    DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks, Engineering Applications of Artificial Intelligence 2023, 123, 106480

    Bayram, F.; Aupke, P.; Ahmed, B.S.; Kassler, A.; Theocharis, A.; Forsman, J. DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks, Engineering Applications of Artificial Intelligence 2023, 123, 106480

    2023

  4. [12]

    Deep Learning in Fault Detection and Diagnosis of building HVAC Systems: A Systematic Review with Meta Analysis, Energy and AI 2023, 12, 100235

    Zhang, F.; Saeed, N.; Sadeghian, P. Deep Learning in Fault Detection and Diagnosis of building HVAC Systems: A Systematic Review with Meta Analysis, Energy and AI 2023, 12, 100235

    2023

  5. [13]

    A systematic review of machine learning techniques related to local energy communities, Renewable and Sustainable Energy Reviews 2022, 170, 112651

    Hernandez-Matheus, A.; Löschenbrand, M.; Berg, K.; Fuchs, I.; Aragüés-Peñalba, M.; Bullich-Massagué, E.; Sumper, A. A systematic review of machine learning techniques related to local energy communities, Renewable and Sustainable Energy Reviews 2022, 170, 112651

    2022

  6. [14]

    An application of the artificial neural network to optimise the energy performances of a magnetic refrigerator, International Journal of Refrigeration 2017, 82, 238--251

    Aprea, C.; Greco, A.; Maiorino, A. An application of the artificial neural network to optimise the energy performances of a magnetic refrigerator, International Journal of Refrigeration 2017, 82, 238--251

    2017

  7. [15]

    Artificial intelligence models for refrigeration, air conditioning and heat pump systems, Energy Reports 2022, 8, 8451--8466

    Adelekan, D.; Ohunakin, O.; Paul, B. Artificial intelligence models for refrigeration, air conditioning and heat pump systems, Energy Reports 2022, 8, 8451--8466

    2022

  8. [16]

    A deep learning framework for building energy consumption forecast, Renewable and Sustainable Energy Reviews 2021, 137, 110591

    Somu, N.; Raman M R , G.; Ramamritham, K. A deep learning framework for building energy consumption forecast, Renewable and Sustainable Energy Reviews 2021, 137, 110591

    2021

  9. [17]

    A review on machine learning forecasting growth trends and their real-time applications in different energy systems, Sustainable Cities and Society 2020, 54, 102010

    Ahmad, T.; Chen, H. A review on machine learning forecasting growth trends and their real-time applications in different energy systems, Sustainable Cities and Society 2020, 54, 102010

    2020

  10. [18]

    Bayesian approach and time series dimensionality reduction to LSTM-based model-building for fault diagnosis of a reciprocating compressor, Neurocomputing 2020, 380, 51--66

    Cabrera, D.; Guamán, A.; Zhang, S.; Cerrada, M.; Sánchez, R.V.; Cevallos, J.; Long, J.; Li, C. Bayesian approach and time series dimensionality reduction to LSTM-based model-building for fault diagnosis of a reciprocating compressor, Neurocomputing 2020, 380, 51--66

    2020

  11. [19]

    An optimized model using LSTM network for demand forecasting, Computers & Industrial Engineering 2020, 143, 106435

    Abbasimehr, H.; Shabani, M.; Yousefi, M. An optimized model using LSTM network for demand forecasting, Computers & Industrial Engineering 2020, 143, 106435

    2020

  12. [20]

    Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: A comprehensive review, Energy and Buildings 2020, 229, 110492

    Mirnaghi, M.S.; Haghighat, F. Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: A comprehensive review, Energy and Buildings 2020, 229, 110492

    2020

  13. [21]

    Large-scale heat pumps: Applications, performance, economic feasibility and industrial integration, Renewable and Sustainable Energy Reviews 2020, 133, 110219

    Schlosser, F.; Jesper, M.; Vogelsang, J.; Walmsley, T.; Arpagaus, C.; Hesselbach, J. Large-scale heat pumps: Applications, performance, economic feasibility and industrial integration, Renewable and Sustainable Energy Reviews 2020, 133, 110219

    2020

  14. [22]

    A long short-term memory artificial neural network to predict daily HVAC consumption in buildings, Energy and Buildings 2020, 216, 109952

    Sendra-Arranz, R.; Guti \'e rrez, A. A long short-term memory artificial neural network to predict daily HVAC consumption in buildings, Energy and Buildings 2020, 216, 109952

    2020

  15. [23]

    Short-term electricity consumption forecast with artificial neural networks—A case study of office buildings, In 2017 IEEE Manchester PowerTech, IEEE, 2017; pp 1--6

    Xypolytou, E.; Meisel, M.; Sauter, T. Short-term electricity consumption forecast with artificial neural networks—A case study of office buildings, In 2017 IEEE Manchester PowerTech, IEEE, 2017; pp 1--6

    2017

  16. [24]

    Using long short-term memory networks to predict energy consumption of air-conditioning systems, Sustainable Cities and Society 2020, 55, 102000

    Zhou, C.; Fang, Z.; Xu, X.; Zhang, X.; Ding, Y.; Jiang, X.; et al. Using long short-term memory networks to predict energy consumption of air-conditioning systems, Sustainable Cities and Society 2020, 55, 102000

    2020

  17. [25]

    A hybrid model approach for forecasting future residential electricity consumption, Energy and Buildings 2016, 117, 341--351

    Dong, B.; Li, Z.; Rahman, S.M.; Vega, R. A hybrid model approach for forecasting future residential electricity consumption, Energy and Buildings 2016, 117, 341--351

    2016

  18. [26]

    Naug, A.; Biswas, G. A data driven method for prediction of energy demand in commercial buildings, In 2018 IEEE 14th International Conference on Automation Science and Engineering (CASE), IEEE, 2018; pp 335--340

    2018

  19. [27]

    Short-term residential load forecasting based on LSTM recurrent neural network, IEEE transactions on smart grid 2017, 10 (1), 841--851

    Kong, W.; Dong, Z.Y.; Jia, Y.; Hill, D.J.; 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

    2017

  20. [28]

    Building thermal load prediction through shallow machine learning and deep learning, Applied Energy 2020, 263, 114683

    Wang, Z.; Hong, T.; Piette, M.A. Building thermal load prediction through shallow machine learning and deep learning, Applied Energy 2020, 263, 114683

    2020

  21. [29]

    A novel energy demand prediction strategy for residential buildings based on ensemble learning, Energy Procedia 2019, 158, 3411--3416

    Huang, Y.; Yuan, Y.; Chen, H.; Wang, J.; Guo, Y.; Ahmad, T. A novel energy demand prediction strategy for residential buildings based on ensemble learning, Energy Procedia 2019, 158, 3411--3416. Innovative Solutions for Energy Transitions

    2019

  22. [30]

    Oil and gold price prediction using optimized fuzzy inference system based extreme learning machine, Resources Policy 2022, 79, 103109

    Das, S.; Sahu, T.P.; Janghel, R.R. Oil and gold price prediction using optimized fuzzy inference system based extreme learning machine, Resources Policy 2022, 79, 103109

    2022

  23. [31]

    Das, S.; Sahu, T.; Janghel, R.; Sahu, B. Effective forecasting of stock market price by using extreme learning machine optimized by PSO-based group oriented crow search algorithm, Neural Computing and Applications 2022, 34

    2022

  24. [32]

    Sahu, R.K.; Shaw, B.; Nayak, J.R.; Shashikant. Short/medium term solar power forecasting of Chhattisgarh state of India using modified TLBO optimized ELM, Engineering Science and Technology, an International Journal 2021, 24 (5), 1180--1200. ://www.sciencedirect.com/science/ar...

    2021

  25. [33]

    Revolutionizing Sustainable Energy Production with Quantum Artificial Intelligence: Applications in Autonomous Robotics and Data Management, Green and Low-Carbon Economy 2023

    Baklaga, L. Revolutionizing Sustainable Energy Production with Quantum Artificial Intelligence: Applications in Autonomous Robotics and Data Management, Green and Low-Carbon Economy 2023

    2023

  26. [34]

    A Semi-Supervised Modulation Identification in MIMO Systems: A Deep Learning Strategy, IEEE Access 2022, 10, 76622--76635

    Bouchenak, S.; Merzougui, R.; Harrou, F.; Dairi, A.; Sun, Y. A Semi-Supervised Modulation Identification in MIMO Systems: A Deep Learning Strategy, IEEE Access 2022, 10, 76622--76635

    2022

  27. [35]

    Automatic fault detection in grid-connected photovoltaic systems via variational autoencoder-based monitoring, Energy Conversion and Management 2024, 314, 118665

    Harrou, F.; Dairi, A.; Taghezouit, B.; Khaldi, B.; Sun, Y. Automatic fault detection in grid-connected photovoltaic systems via variational autoencoder-based monitoring, Energy Conversion and Management 2024, 314, 118665. ://www.sciencedirect.com/science/article/pii/S019689042400606X

    2024

  28. [36]

    TransNAS-TSAD: Harnessing Transformers for Multi-Objective Neural Architecture Search in Time Series Anomaly Detection, Neural Computing & Applications 2024

    Haq, I.U.; Lee, B.S.; Rizzo, D.M. TransNAS-TSAD: Harnessing Transformers for Multi-Objective Neural Architecture Search in Time Series Anomaly Detection, Neural Computing & Applications 2024. Available Online

    2024

  29. [37]

    Peak Anomaly Detection from Environmental Sensor-Generated Watershed Time Series Data, In Information Management and Big Data

    Lee, B.S.; Kaufmann, J.C.; Rizzo, D.M.; Haq, I.U. Peak Anomaly Detection from Environmental Sensor-Generated Watershed Time Series Data, In Information Management and Big Data. SIMBig 2022; Lossio-Ventura, J.A., Valverde-Rebaza, J., Díaz, E., Alatrista-Salas, H., Eds.; Springe...

    2022

  30. [38]

    General review of ground-source heat pump systems for heating and cooling of buildings, Energy and Buildings 2014, 70, 441--454

    Sarbu, I.; Sebarchievici, C. General review of ground-source heat pump systems for heating and cooling of buildings, Energy and Buildings 2014, 70, 441--454

    2014

  31. [39]

    Machine learning-based performance prediction for ground source heat pump systems, Geothermics 2022, 105, 102509

    Zhang, X.; Wang, E.; Liu, L.; Qi, C. Machine learning-based performance prediction for ground source heat pump systems, Geothermics 2022, 105, 102509

    2022

  32. [40]

    Ahmed, N.; Assadi, M.; Ahmed, A.A.; Banihabib, R. Optimal design, operational controls, and data-driven machine learning in sustainable borehole heat exchanger coupled heat pumps: Key implementation challenges and advancement opportunities, Energy for Sustainable Development 2...

    2023

  33. [41]

    RNN-LSTM: From applications to modeling techniques and beyond—Systematic review, Journal of King Saud University - Computer and Information Sciences 2024, 36 (5), 102068

    Al-Selwi, S.M.; Hassan, M.F.; Abdulkadir, S.J.; Muneer, A.; Sumiea, E.H.; Alqushaibi, A.; Ragab, M.G. RNN-LSTM: From applications to modeling techniques and beyond—Systematic review, Journal of King Saud University - Computer and Information Sciences 2024, 36 (5), 102068

    2024

  34. [42]

    A Lightweight Concept Drift Detection and Adaptation Framework for IoT Data Streams, IEEE Internet of Things Magazine 2021, 4 (2), 96--101

    Yang, L.; Shami, A. A Lightweight Concept Drift Detection and Adaptation Framework for IoT Data Streams, IEEE Internet of Things Magazine 2021, 4 (2), 96--101

    2021

  35. [43]

    Light gradient boosting machine with optimized hyperparameters for identification of malicious access in IoT network, Digital Communications and Networks 2023, 9 (1), 125--137

    Mishra, D.; Naik, B.; Nayak, J.; Souri, A.; Dash, P.B.; Vimal, S. Light gradient boosting machine with optimized hyperparameters for identification of malicious access in IoT network, Digital Communications and Networks 2023, 9 (1), 125--137

    2023

  36. [44]

    Intrusion detection model of Internet of Things based on LightGBM, IEICE Transactions on Communications 2023, 106 (8), 622--634

    Zhao, G.; Wang, Y.; Wang, J. Intrusion detection model of Internet of Things based on LightGBM, IEICE Transactions on Communications 2023, 106 (8), 622--634

    2023

  37. [45]

    Multi-sensor Data Fusion Method Based on ARIMA-LightGBM for AGV Positioning, In 2021 5th International Conference on Robotics and Automation Sciences (ICRAS), 2021; pp 272--276

    Che, H. Multi-sensor Data Fusion Method Based on ARIMA-LightGBM for AGV Positioning, In 2021 5th International Conference on Robotics and Automation Sciences (ICRAS), 2021; pp 272--276

    2021

  38. [46]

    Parmezan, A.R.S.; Souza, V.M.; Batista, G.E. Evaluation of statistical and machine learning models for time series prediction: Identifying the state-of-the-art and the best conditions for the use of each model, Information Sciences 2019, 484, 302--337

    2019

  39. [47]

    Long short-term memory, Neural computation 1997, 9 (8), 1735--1780

    Hochreiter, S.; Schmidhuber, J. Long short-term memory, Neural computation 1997, 9 (8), 1735--1780

    1997

  40. [48]

    A review of data-driven approaches for prediction and classification of building energy consumption, Renewable and Sustainable Energy Reviews 2018, 82, 1027--1047

    Wei, Y.; Zhang, X.; Shi, Y.; Xia, L.; Pan, S.; Wu, J.; Han, M.; Zhao, X. A review of data-driven approaches for prediction and classification of building energy consumption, Renewable and Sustainable Energy Reviews 2018, 82, 1027--1047

    2018

  41. [49]

    A comparative performance analysis of different activation functions in LSTM networks for classification, Neural Computing and Applications 2019, 31, 2507--2521

    Farzad, A.; Mashayekhi, H.; Hassanpour, H. A comparative performance analysis of different activation functions in LSTM networks for classification, Neural Computing and Applications 2019, 31, 2507--2521

    2019

  42. [50]

    LSTM inefficiency in long-term dependencies regression problems, Journal of Advanced Research in Applied Sciences and Engineering Technology 2023, 30 (3), 16--31

    Al-Selwi, S.M.; Hassan, M.F.; Abdulkadir, S.J.; Muneer, A. LSTM inefficiency in long-term dependencies regression problems, Journal of Advanced Research in Applied Sciences and Engineering Technology 2023, 30 (3), 16--31

    2023

  43. [51]

    A survey on anomaly detection for technical systems using LSTM networks, Computers in Industry 2021, 131, 103498

    Lindemann, B.; Maschler, B.; Sahlab, N.; Weyrich, M. A survey on anomaly detection for technical systems using LSTM networks, Computers in Industry 2021, 131, 103498. ://www.sciencedirect.com/science/article/pii/S0166361521001056

    2021

  44. [52]

    A novel attLSTM framework combining the attention mechanism and bidirectional LSTM for demand forecasting, Expert Systems with Applications 2024, 254, 124409

    Cui, L.; Chen, Y.; Deng, J.; Han, Z. A novel attLSTM framework combining the attention mechanism and bidirectional LSTM for demand forecasting, Expert Systems with Applications 2024, 254, 124409

    2024

  45. [53]

    Paraphrase detection using LSTM networks and handcrafted features, Multimedia Tools and Applications 2021, 80 (4), 6479--6492

    Shahmohammadi, H.; Dezfoulian, M.; Mansoorizadeh, M. Paraphrase detection using LSTM networks and handcrafted features, Multimedia Tools and Applications 2021, 80 (4), 6479--6492

    2021

  46. [54]

    Zhou, P.; Shi, W.; Tian, J.; Qi, Z.; Li, B.; Hao, H.; Xu, B. Attention-based bidirectional long short-term memory networks for relation classification, In Proceedings of the 54th annual meeting of the association for computational linguistics (volume 2: Short papers), 2016; pp...

    2016

  47. [55]

    Feng, T.; Zhao, Y.; Shi, H.; Li, H.; Zhang, X.; Wang, G.; Badawi, R.D.; Price, P.M.; Jones, T.; Cherry, S.R. The effects of delay on the input function for early dynamics in total body parametric imaging, In 2019 IEEE Nuclear Science Symposium and Medical Imaging Conference (N...

    2019

  48. [56]

    Unidirectional and bidirectional LSTM models for short-term traffic prediction, Journal of Advanced Transportation 2021, 2021 (1), 5589075

    Abduljabbar, R.L.; Dia, H.; Tsai, P.W. Unidirectional and bidirectional LSTM models for short-term traffic prediction, Journal of Advanced Transportation 2021, 2021 (1), 5589075

    2021

  49. [57]

    Extended Isolation Forest, IEEE Transactions on Knowledge and Data Engineering 2021, 33 (4), 1479--1489

    Hariri, S.; Kind, M.C.; Brunner, R.J. Extended Isolation Forest, IEEE Transactions on Knowledge and Data Engineering 2021, 33 (4), 1479--1489

    2021

  50. [58]

    Anomaly detection: A survey, ACM Comput

    Chandola, V.; Banerjee, A.; Kumar, V. Anomaly detection: A survey, ACM Comput. Surv. 2009, 41 (3)

    2009

  51. [59]

    New Trends in Time Series Anomaly Detection., In EDBT, 2023; pp 847--850

    Boniol, P.; Paparizzos, J.; Palpanas, T. New Trends in Time Series Anomaly Detection., In EDBT, 2023; pp 847--850

    2023

  52. [60]

    An In-Depth Study and Improvement of Isolation Forest, IEEE Access 2022, 10, 10219--10237

    Chabchoub, Y.; Togbe, M.U.; Boly, A.; Chiky, R. An In-Depth Study and Improvement of Isolation Forest, IEEE Access 2022, 10, 10219--10237

    2022

  53. [61]

    l-DBSCAN : A Fast Hybrid Density Based Clustering Method, In 18th International Conference on Pattern Recognition (ICPR'06), 2006; Vol

    Viswanath, P.; Pinkesh, R. l-DBSCAN : A Fast Hybrid Density Based Clustering Method, In 18th International Conference on Pattern Recognition (ICPR'06), 2006; Vol. 1, pp 912--915

    2006

  54. [62]

    Approximate training of one-class support vector machines using expected margin, Computers & Industrial Engineering 2019, 130, 772--778

    Kang, S.; Kim, D.; Cho, S. Approximate training of one-class support vector machines using expected margin, Computers & Industrial Engineering 2019, 130, 772--778

    2019

  55. [63]

    Autoencoders and their applications in machine learning: a survey, Artificial Intelligence Review 2024, 57 (2), 28

    Berahmand, K.; Daneshfar, F.; Salehi, E.S.; Li, Y.; Xu, Y. Autoencoders and their applications in machine learning: a survey, Artificial Intelligence Review 2024, 57 (2), 28

    2024

  56. [64]

    Isolation forest, In 2008 eighth ieee international conference on data mining, IEEE, 2008; pp 413--422

    Liu, F.T.; Ting, K.M.; Zhou, Z.H. Isolation forest, In 2008 eighth ieee international conference on data mining, IEEE, 2008; pp 413--422

    2008

  57. [65]

    Vitorino, J.; Andrade, R.; Pra c a, I.; Sousa, O.; Maia, E. A Comparative Analysis of Machine Learning Techniques for IoT Intrusion Detection, In Foundations and Practice of Security, Cham; A \"i meur, E., Laurent, M., Yaich, R., Dupont, B., Garcia-Alfaro, J., Eds.; Springer I...

    2022

  58. [66]

    Effects of data preprocessing on detecting autism in adults using web-based eye-tracking data, Behaviour & Information Technology 2023, 42 (14), 2476--2484

    Erfan Khalaji, Y.Y., Sukru Eraslan; Yaneva, V. Effects of data preprocessing on detecting autism in adults using web-based eye-tracking data, Behaviour & Information Technology 2023, 42 (14), 2476--2484

    2023

  59. [67]

    Missing data in time series: A review of imputation methods and case study, Learning and Nonlinear Models 2022, 20 (1), 31--46

    Ribeiro, S.M.; Castro, C. Missing data in time series: A review of imputation methods and case study, Learning and Nonlinear Models 2022, 20 (1), 31--46

    2022

  60. [68]

    v‐plots: Designing Hybrid Charts for the Comparative Analysis of Data Distributions, Computer Graphics Forum 2020, 39, 565--577

    Blumenschein, M.; Debbeler, L.; Lages, N.; Renner, B.; Keim, D.; El-Assady, M. v‐plots: Designing Hybrid Charts for the Comparative Analysis of Data Distributions, Computer Graphics Forum 2020, 39, 565--577

    2020

  61. [69]

    Random feature selection using random subspace logistic regression, Expert Systems with Applications 2023, 217, 119535

    Wichitaksorn, N.; Kang, Y.; Zhang, F. Random feature selection using random subspace logistic regression, Expert Systems with Applications 2023, 217, 119535

    2023

  62. [70]

    Anomaly detection based on machine learning in IoT-based vertical plant wall for indoor climate control, Building and Environment 2020, 183, 107212

    Liu, Y.; Pang, Z.; Karlsson, M.; Gong, S. Anomaly detection based on machine learning in IoT-based vertical plant wall for indoor climate control, Building and Environment 2020, 183, 107212

    2020

  63. [71]

    Electricity load forecasting: a systematic review, Journal of Electrical Systems and Information Technology 2020, 7, 1--19

    Nti, I.K.; Teimeh, M.; Nyarko-Boateng, O.; Adekoya, A.F. Electricity load forecasting: a systematic review, Journal of Electrical Systems and Information Technology 2020, 7, 1--19

    2020

  64. [72]

    Global vs

    Buonanno, A.; Caliano, M.; Pontecorvo, A.; Sforza, G.; Valenti, M.; Graditi, G. Global vs. Local Models for Short-Term Electricity Demand Prediction in a Residential/Lodging Scenario, Energies 2022, 15 (6)

    2022

  65. [73]

    , " * write output.state after.block =

    ENTRY address archive author booktitle chapter collaboration edition editor eid howpublished institution journal key lastchecked month note number numpages organization pages publisher school series title type url volume year label INTEGERS output.state before.all mid.sentence...

  66. [74]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.