A Systematic Literature Review on Machine Learning Techniques for Predicting Household Water Consumption

Supplementary Material 1 Detailed Methodological Workflow and Execution Timeline The authors employed appropriate visualization tools to support a clear understanding and effective presentation of the data obtained from the selected studies. Specifically, tables were used to organize and display the extracted information in direct relation to the research questions. Additionally, a bar chart was included to illustrate the annual distribution of the primary studies within the selected time frame. These visual representations facilitate the identification of trends and ensure alignment with the standards of a systematic literature review that focuses on machine learning techniques for household water consumption prediction. To implement the seven steps of this methodology, the authors employed a structured approach. The authors adopted a collaborative workflow to ensure consistency. Step 1-4. (Planning-Formulation-Search-Criteria). Using Parsifal1, Web application to assist the execution of an SLR, authors could carry out the first four steps. Parsifal allowed authors to define and execute the string queries in the reference databases: Wiley, ScienceDirect, Scopus, SpringerLink, ACM, IEEE, and WoS. At the end of Step 3, the author team verified and collected the execution results of each reference database. To apply the criteria of inclusion and exclusion of Step 4, Parsifal provides a User Interface that allowed us to assist and administer when we selected included,excluded, and maybe for each paper. Time: These steps were completed over a period of one month. Step 5. (Quality). This step took longer than the others because the remaining papers from the previous steps were reviewed carefully. Using a shared Google Sheet file2, each author assigned a value (quality assessment) to a subset of all the collections of papers. When one author’s paper assessment was not clear, the rest of the authors carried out the assessment and assigned the average of the authors’ values. Time: The quality assessment phase required approximately three months. Step 6-7. (Collection-Analysis). The authors consolidated the results from the previous steps into a uniform dataset. Subsequently, the team analyzed this data to generate a comprehensive set of visualizations, selecting the most representative charts and tables for this study. Time: The entire process spanned approximately two months: two weeks for data consolidation (Step 6) and 1.5 months for analysis and visualization (Step 7). Summary. We implemented the seven methodological steps through a structured, equitable workflow among all authors. Steps 1-4 (planning, research questions, search, and selection) were executed with Parsifal, which managed database queries and include,exclude, or maybe decisions, and together took one month. For Step 5 (quality assessment), authors independently scored papers in a shared Google Sheet; unclear cases were jointly reviewed and resolved by averaging scores, a phase that lasted three months. Steps 6-7 produced a unified dataset and the final set of charts/tables, with data collection taking half a month and analysis/selection taking 1.5 months, ensuring transparency and methodological rigor. 1https://parsif.al. 2https://doi.org/10.5281/zenodo.18018855 1 2 Primary Studies Table 1: Primary studies: ID, reference, and publication type (Journal or Conference) ID Reference Type ID Reference Type P01 [Niknam et al., 2023a] Journal P41 [Uzlu, 2024] Journal P02 [Pu et al., 2022] Journal P42 [Gautam et al., 2020] Journal P03 [Michalopoulos et al., 2024] Journal P43 [Dai et al., 2011] Conference P04 [Zubaidi et al., 2020b] Journal P44 [Guo et al., 2022] Journal P05 [Jithish and Sankaran, 2017] Conference P45 [Felfelani and and, 2016] Journal P06 [Zubaidi et al., 2020a] Journal P46 [Velasco et al., 2018] Conference P07 [Sajadifar and Pakseresht, 2024] Journal P47 [Vo et al., 2022] Conference P08 [Nasser et al., 2020] Journal P48 [Niyongabo et al., 2024] Journal P09 [Zheng et al., 2022] Journal P49 [Pakpahan et al., 2023] Conference P10 [Suh and Ham, 2016] Journal P50 [Li et al., 2023] Journal P11 [Rustam et al., 2022] Journal P51 [Zhang and Wang, 2022] Conference P12 [Karamaziotis et al., 2020] Journal P52 [Cao et al., 2024] Journal P13 [Sebri, 2013] Journal P53 [Wang et al., 2023] Journal P14 [Zhou et al., 2024] Journal P54 [Kavya et al., 2023] Journal P15 [Smolak et al., 2020] Journal P55 [Zhang et al., 2022] Journal P16 [Nguyen et al., 2016] Conference P56 [Xenochristou et al., 2021] Journal P17 [Aggarwal and Sehgal, 2022] Conference P57 [Alsumaiei, 2021] Journal P18 [Görenekli and Gülbağ, 2024] Journal P58 [Namdari et al., 2023] Journal P19 [Ghannam and Hussain, 2023] Journal P59 [Thakur et al., 2021] Journal P20 [Deng et al., 2022] Conference P60 [Athapaththu et al., 2020] Conference P21 [Li and Fu, 2024] Journal P61 [Polić et al., 2023] Conference P22 [Zubaidi et al., 2023] Journal P62 [Menapace et al., 2021] Journal P23 [Niknam et al., 2023b] Journal P63 [Candelieri et al., 2019] Journal P24 [Sajjanshetty et al., 2023] Conference P64 [Said et al., 2021] Journal P25 [Pourmousavi et al., 2022] Journal P65 [Shi et al., 2013] Journal P26 [Yin and Xiong, 2024] Journal P66 [Oyebode and Ighravwe, 2019] Journal P27 [El Hanjri et al., 2023] Journal P67 [Yang et al., 2023] Journal P28 [Chatzigeorgakidis et al., 2018] Journal P68 [Kesornsit and Sirisathitkul, 2022] Journal P29 [Cao et al., 2023] Journal P69 [Drevetskyi et al., 2018] Conference P30 [Jiang et al., 2024] Journal P70 [Yan et al., 2022] Conference P31 [Mumbi et al., 2022] Journal P71 [García-Soto et al., 2024] Journal P32 [Cheng et al., 2023] Conference P72 [dos Santos and Pereira Filho, 2014] Journal P33 [Li et al., 2024] Journal P73 [Chang and Liu, 2009] Conference P34 [Iwakin and Moazeni, 2024] Journal P74 [Yurdusev et al., 2009] Journal P35 [Abu Talib et al., 2023] Journal P75 [Faiz and Daniel, 2022] Conference P36 [Bashar et al., 2023] Journal P76 [Compagnon et al., 2022] Conference P37 [Ndayisenga et al., 2022] Conference P77 [Wang et al., 2024] Conference P38 [Filho et al., 2024] Journal P78 [Sahoo et al., 2023] Journal P39 [Santos de Jesus and Silva Gomes, 2023] Journal P79 [Xu, 2024] Journal P40 [Gil-Gamboa et al., 2024] Journal P80 [Kim et al., 2022] Journal 2 3 Detailed Characterization of Machine Learning Model Properties (RQ1) This supplementary section provides additional descriptive detail supporting the synthesis presented in Section 3.3 of the main manuscript. The purpose of this material is to document the conceptual classification used to analyze machine learning (ML) model characteristics, while preserving conciseness in the main text. All quantitative mappings associated with this analysis are provided in the accompanying Excel evidence matrix available in the public repository. To address Research Question 1 (RQ1), each of the 80 primary studies was systematically reviewed to extract the characteristics attributed to the ML techniques applied in residential water consumption forecasting. Because substantial heterogeneity in terminology was observed across studies, the extracted attributes were consolidated into 20 high level conceptual categories. This consolidation was performed to reduce semantic variability while preserving the methodological traits emphasized by the original authors. The overall distribution of these characteristics across the reviewed studies is summarized in the corresponding figures presented in Section 3.3 of the main manuscript. As reported there, nonlinear modeling capability, reliability, and long term dependency modeling are the most frequently cited properties, reflecting the need to capture complex temporal dynamics and interacting behavioral and environmental factors in household water consumption. For completeness, the conceptual categories used in the classification are described below. Nonlinear modeling capability. The capacity of a model to learn complex nonlinear relationships between variables, which is essential in residential water consumption contexts where environmental conditions and human behavior interact in non trivial ways. Reliability. The ability of a model to produce stable and consistent results across different datasets, temporal horizons, or operational conditions. Long term dependency modeling. The capacity to capture relationships between temporally distant observations, which is particularly relevant for medium and long horizon forecasting. Improved version of another model. Enhancements to existing approaches through architectural modifications, parameter optimization, or hybridization aimed at improving predictive performance. Good handling of time series. Suitability for sequential data, including the representation of trends, seasonality, and temporal correlations. Hybrid model. Architectures that combine multiple modeling paradigms in order to leverage complementary strengths and improve robustness or generalization. Popular model. Techniques that are widely adopted in the literature, often due to established effectiveness and ease of implementation. Applied in real world scenarios. Models validated using real consumption data or operational settings beyond purely academic benchmarks. Fast training. Computational efficiency during the training phase, particularly when applied to large datasets. Effective feature extraction. The ability to automatically identify informative patterns or representations from raw input data, reducing reliance on manual feature engineering. Simplicity. Preference for minimal architectural complexity, often associated with ease of implementation and transparency. Benchmark model. Techniques commonly used as reference baselines for comparative evaluation. Error driven learning. Learning mechanisms in which internal parameters are updated directly in response to prediction errors. Complex pattern learning. The capacity to capture intricate or hidden structures in data, such as nonlinear interactions or hierarchical dependencies. Interpretable. Models designed to offer insight into the decision making process, either intrinsically or through post hoc explanation techniques. Elegant design. Architectures that balance simplicity and effectiveness through well integrated components. Robustness. The ability to maintain performance in the presence of noise, missing values, or unexpected perturbations. Complex model. Architectures involving multiple layers, nonlinear transformations, or advanced mechanisms such as attention or memory units. Ensemble learning. Techniques that aggregate multiple learners in order to improve predictive accuracy and reduce overfitting. Temporal trends in both model adoption and attributed characteristics are analyzed and visualized in the main manuscript. As discussed there, a pronounced shift occurs around 2020, characterized by increased adoption of deep learning based architectures and a corresponding rise in characteristics related to nonlinear modeling and long term dependency capture. This alignment supports the conclusion that recent methodological choices are driven primarily by the need to represent complex temporal dynamics inherent to residential water consumption. This supplementary analysis reinforces the findings reported in the main manuscript. The selection of ML techniques is predominantly guided by the ability to model nonlinear relationships, ensure reliable performance, and capture long term temporal dependencies. These priorities reflect the multifactorial and dynamic nature of household water demand and provide context for the taxonomy and synthesis presented in the primary results section. 3 4 Reported Benefits Associated with Evaluation Metrics (RQ3) This supplementary section provides additional detail supporting the analysis of performance metrics presented in Section 3.5 of the main manuscript. The focus of this material is on the benefits reported by authors in association with the evaluation of machine learning (ML) techniques for residential water consumption forecasting. These descriptive findings were moved to the supplementary material to preserve conciseness in the Results section while maintaining transparency. To complement the identification of dominant evaluation metrics, the benefits explicitly reported in the reviewed studies were systematically extracted and categorized. Each benefit was associated with the evaluation context in which it was discussed, resulting in a set of qualitative descriptors reflecting how authors interpreted model performance. The complete mapping between studies, metrics, and reported benefits is provided in the accompanying Excel evidence matrix. Figure 1 summarizes the frequency with which different benefits were cited across the reviewed studies. Precision, reliability, and performance were identified as the most frequently reported benefits, indicating that evaluation practices are primarily oriented toward minimizing prediction error and ensuring stable model behavior. Benefits related to interpretability, scalability, and computational efficiency were reported less frequently, despite their relevance for operational deployment. Figure 1: Reported benefits of machine learning techniques as described in the reviewed studies. Precision, reliability, and performance are the most frequently cited benefits Precision-related benefits were commonly associated with magnitude-based error metrics such as Root Mean Squared Error and Mean Absolute Error, which are frequently used to assess short-term predictive accuracy. Reliability was often reported in studies emphasizing consistency of performance across multiple datasets or temporal segments, particularly in applications involving recurrent or ensemble-based models. Performance-related benefits were generally described in broad terms, encompassing improvements in predictive accuracy, convergence behavior, or robustness relative to baseline approaches. Less frequently cited benefits, including interpretability and computational efficiency, were typically mentioned in studies employing simpler model architectures or post hoc explanation techniques. The limited emphasis on these aspects suggests that evaluation practices remain largely focused on predictive accuracy rather than on properties related to transparency, scalability, or resource constraints. The findings reported in this supplementary section reinforce the conclusions presented in the main manuscript for RQ3. Evaluation practices in residential water consumption forecasting are dominated by accuracy-oriented metrics and benefits, while other dimensions relevant to deployment and governance receive comparatively limited attention. This imbalance provides additional context for the discussion of evaluation standardization and methodological limitations addressed in subsequent sections. 4 5 Citation per year We complement the descriptive synthesis with a lightweight citation analysis to contextualize influence within the corpus. Using a single indexing source and a fixed retrieval date for all papers, we recorded total citations and computed citations per year (CPY) to mitigate age effects. For each study i: CPYi=Citesi ∆ti ,(1a) ∆ti= max1,tretrieval −tpublication 365.25 .(1b) This definition enforces a minimum exposure window of one year for recently published articles. When multiple versions existed (e.g., preprint and journal), we used the peer-reviewed canonical version and consolidated citation counts to avoid duplication. Each study was then mapped to taxonomy labels M1–M6 and P1–P7 following the evidence-aligned rules in Model–Predictor Crosswalk Section; hybrids were coded as M6 whenever a decomposition or statistical component was explicitly combined with machine learning on residuals. Because some studies evaluate more than one model family, they may appear in multiple family blocks. This reflects influence within each research stream rather than implying additivity across families. Table 2: Top 5 Most influential primary studies by Total Cites ID Study reference Year Cites CPY P04 [Zubaidi et al., 2020b] 2020 149 29.80 P06 [Zubaidi et al., 2020a] 2020 134 26.80 P63 [Candelieri et al., 2019] 2018 97 13.86 P08 [Nasser et al., 2020] 2020 75 15.00 P15 [Smolak et al., 2020] 2020 74 14.80 Table 3: Top 5 Most influential primary studies by Cites Per Year (CPY) ID Study reference Year Cites CPY P04 [Zubaidi et al., 2020b] 2020 149 29.80 P06 [Zubaidi et al., 2020a] 2020 134 26.80 P02 [Pu et al., 2022] 2023 46 23.00 P54 [Kavya et al., 2023] 2023 44 22.00 P08 [Nasser et al., 2020] 2020 75 15.00 Table 4: Top studies by model family by citations Model family Predictor families ID Study reference Cites CPY M1 P1 P04 [Zubaidi et al., 2020b] 149 29.80 M1 P3 P06 [Zubaidi et al., 2020a] 134 26.80 M1 P2, P4, P6 P08 [Nasser et al., 2020] 75 15.00 M2 P2, P4, P6 P08 [Nasser et al., 2020] 75 15.00 M2 P4 P15 [Smolak et al., 2020] 74 14.80 M2 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M3 P3 P63 [Candelieri et al., 2019] 97 13.86 M3 P2, P4, P6 P08 [Nasser et al., 2020] 75 15.00 M3 P4 P15 [Smolak et al., 2020] 74 14.80 M4 P1, P3 P72 [dos Santos and Pereira Filho, 2014] 56 5.09 M4 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M4 P2 P13 [Sebri, 2016] 17 1.42 M5 P3, P4 P28 [Chatzigeorgakidis et al., 2018] 45 6.43 M5 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M5 P1, P4, P5, P6 P25 [Pourmousavi et al., 2022] 8 2.67 M6 P1, P2, P3 P02 [Pu et al., 2022] 46 23.00 M6 P1 P44 [Guo et al., 2022] 30 10.00 M6 P2 P78 [Sahoo et al., 2023] 18 9.00 Tables 2–5 summarize the results. By total citations (Table 2), two 2020 works dominate, with additional strong entries from 2018 and 2020. When normalized by year (Table 3), recent contributions surface with steep annual accrual. Stratifying by model family (Tables 4 and 5) reinforces two strands aligned with our crosswalk: M1 paired with P2,P1,P3 and M2 with P6. Kernel methods M3 are influential where nonlinearity is modest and the feature space is compact, linear baselines M4 are prominent in small Nor policy-interpretability settings, and hybrids M6 show rising CPY where seasonality or regime changes are explicit. This pattern coheres with the model–predictor crosswalk and the decision tree for model selection. Two caveats are important. First, citation counts depend on source coverage (journals versus preprints, conference proceedings) and the retrieval date; CPY reduces but does not eliminate recency effects. Second, influence is not performance: highly cited studies may be early, comprehensive, or widely accessible rather than uniformly superior on error metrics. We therefore use the tables to contextualize thematic prominence and to identify anchor papers within each family, not to rank algorithms. In practice, treat the leaders in each family as starting points for replication and ablation (e.g., reproduce feature 5 Table 5: Top studies by model family by CPY Model family Predictor families ID Study reference Cites CPY M1 P1 P04 [Zubaidi et al., 2020b] 149 29.80 M1 P3 P06 [Zubaidi et al., 2020b] 134 26.80 M1 P1, P2, P3 P02 [Pu et al., 2022] 46 23.00 M2 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M2 P2, P4, P6 P08 [Nasser et al., 2020] 75 15.00 M2 P4 P15 [Smolak et al., 2020] 74 14.80 M3 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M3 P2, P4, P6 P08 [Nasser et al., 2020] 75 15.00 M3 P4 P15 [Smolak et al., 2020] 74 14.80 M4 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M4 P1, P2, P4 P34 [Iwakin and Moazeni, 2024] 9 9.00 M4 P1, P3 P72 [dos Santos and Pereira Filho, 2014] 56 5.09 M5 P1, P2, P3 P54 [Kavya et al., 2023] 44 22.00 M5 P3, P4 P28 [Chatzigeorgakidis et al., 2018] 45 6.43 M5 P1, P4, P5, P6 P25 [Pourmousavi et al., 2022] 8 2.67 M6 P1, P2, P3 P02 [Pu et al., 2022] 46 23.00 M6 P1 P44 [Guo et al., 2022] 30 10.00 M6 P2 P78 [Sahoo et al., 2023] 18 9.00 sets by P1–P7, validate with blocked or rolling-origin cross-validation, and report both a scale-dependent error such as RMSE or MAE and a scale-free metric such as sMAPE or MASE). References [Abu Talib et al., 2023] Abu Talib, M., Abdallah, M., Abdeljaber, A., and Abu Waraga, O. (2023). Influence of exogenous factors on water demand forecasting models during the COVID-19 period. Engineering Applications of Artificial Intelligence, 117:105617. [Aggarwal and Sehgal, 2022] Aggarwal, S. and Sehgal, S. (2022). Automation of Water Distribution System by Prediction of Water Consumption and Leakage Detection Using Machine Learning and IoT. In 2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), pages 345–350. [Alsumaiei, 2021] Alsumaiei, A. A. (2021). Short-term forecasting of monthly water consumption in hyper-arid climate using recurrent neural networks. Journal of Engineering Research, 9(3B):56–69. [Athapaththu et al., 2020] Athapaththu, A. M. H. N., Illeperumarachchi, D. U. S., Herath, H. M. K. U., Jayasinghe, H. K., Rankothge, W. H., and Gamage, N. (2020). Supply and Demand Planning for Water: A Sustainable Water Management System. In 2020 2nd International Conference on Advancements in Computing (ICAC), pages 305–310. [Bashar et al., 2023] Bashar, A. M., Nozari, H., Marofi, S., Mohamadi, M., and Ahadiiman, A. (2023). Investigation of factors affecting rural drinking water consumption using intelligent hybrid models. Water Science and Engineering, 16(2):175–183. [Candelieri et al., 2019] Candelieri, A., Giordani, I., Archetti, F., Barkalov, K., Meyerov, I., Polovinkin, A., Sysoyev, A., and Zolotykh, N. (2019). Tuning hyperparameters of a SVM-based water demand forecasting system through parallel global optimization. Computers & Operations Research, 106:202–209. [Cao et al., 2023] Cao, L., Yuan, X., Tian, F., Xu, H., and Su, Z. (2023). Forecasting of water consumption by integrating spatial and temporal characteristics of short-term water use in cities. Physics and Chemistry of the Earth, Parts A/B/C, 130:103390. [Cao et al., 2024] Cao, Y., Wang, Z., Li, P., Zhou, Z., Li, W., Zheng, T., Liu, J., Wu, W., Shi, Z., and Liu, J. (2024). Prediction of rural domestic water and sewage production based on automated machine learning in northern China. Journal of Cleaner Production, 434:140016. [Chang and Liu, 2009] Chang, M. and Liu, J. (2009). Water Demand Prediction Model Based on Radial Basis Function Neural Network. In 2009 First International Conference on Information Science and Engineering, pages 5295–5298. [Chatzigeorgakidis et al., 2018] Chatzigeorgakidis, G., Karagiorgou, S., Athanasiou, S., and Skiadopoulos, S. (2018). FMLkNN: scalable machine learning on Big Data using k-nearest neighbor joins. Journal of Big Data, 5(1):4. [Cheng et al., 2023] Cheng, Y., Qian, Z., and Bao, K. (2023). Hot Water Load Prediction Based on a Combined Model of CNN-GRU and TCN. In 2023 4th International Conference on Advanced Electrical and Energy Systems (AEES), pages 300–304. [Compagnon et al., 2022] Compagnon, P., Lomet, A., Reyboz, M., and Mermillod, M. (2022). Domestic hot water forecasting for individual housing with deep learning. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pages 223–235. 6 [Dai et al., 2011] Dai, J., Li, M., Sahu, S., Naphade, M., and Chen, F. (2011). Multi-granular demand forecasting in SmarterWater. In Proceedings of the 13th International Conference on Ubiquitous Computing, UbiComp ’11, page 595–596, New York, NY, USA. Association for Computing Machinery. [Deng et al., 2022] Deng, L., Chang, X., and Wang, P. (2022). Daily Water Demand Prediction Driven by Multi-source Data. Procedia Computer Science, 208:128–135. 7th International Conference on Intelligent, Interactive Systems and Applications. [dos Santos and Pereira Filho, 2014] dos Santos, C. C. and Pereira Filho, A. J. (2014). Water Demand Forecasting Model for the Metropolitan Area of São Paulo, Brazil. Water Resources Management, 28(13):4401–4414. [Drevetskyi et al., 2018] Drevetskyi, V., Klepach, M., and Kutia, V. (2018). Water Consumption Prediction for City Pumping Station Using Neural Networks. In Burduk, A. and Mazurkiewicz, D., editors, Intelligent Systems in Production Engineering and Maintenance – ISPEM 2017, pages 459–467, Cham. Springer International Publishing. [El Hanjri et al., 2023] El Hanjri, M., Kabbaj, H., Kobbane, A., and Abouaomar, A. (2023). Federated learning for water consumption forecasting in smart cities. In ICC 2023-IEEE International Conference On Communications, pages 1798– 1803. IEEE, Institute of Electrical and Electronics Engineers Inc. [Faiz and Daniel, 2022] Faiz, M. and Daniel, A. K. (2022). Wireless Sensor Network Based Distribution and Prediction of Water Consumption in Residential Houses Using ANN. In Misra, R., Kesswani, N., Rajarajan, M., Veeravalli, B., and Patel, A., editors, Internet of Things and Connected Technologies, pages 107–116, Cham. Springer International Publishing. [Felfelani and and, 2016] Felfelani, F. and and, R. K. (2016). Municipal water demand forecasting under peculiar fluctuations in population: a case study of Mashhad, a tourist city. Hydrological Sciences Journal, 61(8):1524–1534. [Filho et al., 2024] Filho, J. V., Scortegagna, A., Vieira, A. P. d. S. D., and Jaskowiak, P. A. (2024). Machine learning for water demand forecasting: Case study in a Brazilian coastal city. Water Practice and Technology, 19(5):1586–1602. [García-Soto et al., 2024] García-Soto, C. G., Torres, J. F., Zamora-Izquierdo, M. A., Palma, J., and Troncoso, A. (2024). Water consumption time series forecasting in urban centers using deep neural networks. Applied Water Science, 14(2):21. [Gautam et al., 2020] Gautam, J., Chakrabarti, A., Agarwal, S., Singh, A., Gupta, S., and Singh, J. (2020). Monitoring and forecasting water consumption and detecting leakage using an IoT system. Water Supply, 20(3):1103–1113. [Ghannam and Hussain, 2023] Ghannam, S. and Hussain, F. (2023). Comparison of deep learning approaches for forecasting urban short-term water demand a Greater Sydney Region case study. Knowledge-Based Systems, 275:110660. [Gil-Gamboa et al., 2024] Gil-Gamboa, A., Paneque, P., Trull, O., and Troncoso, A. (2024). Medium-term water consumption forecasting based on deep neural networks. Expert Systems with Applications, 247:123234. [Guo et al., 2022] Guo, J., Sun, H., and Du, B. (2022). Multivariable Time Series Forecasting for Urban Water Demand Based on Temporal Convolutional Network Combining Random Forest Feature Selection and Discrete Wavelet Transform. Water Resources Management, 36(9):3385–3400. [Görenekli and Gülbağ, 2024] Görenekli, K. and Gülbağ, A. (2024). Comparative Analysis of Machine Learning Techniques for Water Consumption Prediction: A Case Study from Kocaeli Province. Sensors, 24(17). [Iwakin and Moazeni, 2024] Iwakin, O. and Moazeni, F. (2024). Improving urban water demand forecast using conformal prediction-based hybrid machine learning models. Journal of Water Process Engineering, 58:104721. [Jiang et al., 2024] Jiang, Q., Guo, W., Wang, Z., Wu, Y., Zhao, Y., Tao, M., and Sun, Y. (2024). Forecasting regional water demand using multi-fidelity data and harris hawks optimization of generalized regression neural network models – A case study of Heilongjiang Province, China. Journal of Hydrology, 634:131084. [Jithish and Sankaran, 2017] Jithish, J. and Sankaran, S. (2017). A Neuro-Fuzzy Approach for Domestic Water Usage Prediction. In 2017 IEEE Region 10 Symposium (TENSYMP), pages 1–5. [Karamaziotis et al., 2020] Karamaziotis, P. I., Raptis, A., Nikolopoulos, K., Litsiou, K., and Assimakopoulos, V. (2020). An empirical investigation of water consumption forecasting methods. International Journal of Forecasting, 36:588–606. [Kavya et al., 2023] Kavya, M., Mathew, A., Shekar, P. R., and P, S. (2023). Short term water demand forecast modelling using artificial intelligence for smart water management. Sustainable Cities and Society, 95:104610. [Kesornsit and Sirisathitkul, 2022] Kesornsit, W. and Sirisathitkul, Y. (2022). Water consumption prediction based on machine learning methods and public data. Advances in Computational Design, 7(2):113–128. 7 [Kim et al., 2022] Kim, J., Lee, H., Lee, M., Han, H., Kim, D., and Kim, H. S. (2022). Development of a deep learning-based prediction model for water consumption at the household level. Water, 14(9):1512. [Li and Fu, 2024] Li, D. and Fu, Q. (2024). Deep Learning Model-Based Demand Forecasting for Secondary Water Supply in Residential Communities: A Case Study of Shanghai City, China. IEEE Access, 12:38745–38757. [Li et al., 2023] Li, Z., Peng, S., Zheng, G., Chu, X., and Tian, Y. (2023). Prediction of Daily Water Consumption in Residential Areas Based on Meteorologic Conditions—Applying Gradient Boosting Regression Tree Algorithm. Water, 15(19). [Li et al., 2024] Li, Z., Wang, G., Lin, D., and Mashhadi, A. (2024). Hybrid approach for accurate water demand prediction using socio-economic and climatic factors with ELM optimization. Heliyon, 10(3):e25028. [Menapace et al., 2021] Menapace, A., Zanfei, A., and Righetti, M. (2021). Tuning ANN Hyperparameters for Forecasting Drinking Water Demand. Applied Sciences, 11(9). [Michalopoulos et al., 2024] Michalopoulos, C., Dimas, P., Kossieris, P., Pelekanos, N., and Makropoulos, C. (2024). A largescale evaluation of machine learning algorithms in mid-term water demand forecasting. Water Practice and Technology, 19(7):2693–2711. [Mumbi et al., 2022] Mumbi, A. W., Li, F., Bavumiragira, J. P., and Fangninou, F. F. (2022). Forecasting water consumption on transboundary water resources for water resource management using the feed-forward neural network: a case study of the Nile River in Egypt and Kenya. Marine and Freshwater Research, 73(3):292–306. [Namdari et al., 2023] Namdari, H., Haghighi, A., and Ashrafi, S. M. (2023). Short-term urban water demand forecasting; application of 1D convolutional neural network (1D CNN) in comparison with different deep learning schemes. Stochastic Environmental Research and Risk Assessment. [Nasser et al., 2020] Nasser, A. A., Rashad, M. Z., and Hussein, S. E. (2020). A Two-Layer Water Demand Prediction System in Urban Areas Based on Micro-Services and LSTM Neural Networks. IEEE Access, 8:147647–147661. [Ndayisenga et al., 2022] Ndayisenga, G., Gatera, O., Kabiri, C., Niyitegeka, J., Bampire, D., and Harerimana, S. (2022). IoT-Based Household Water Consumption Management System. In 2022 IEEE PES/IAS PowerAfrica, pages 1–4. [Nguyen et al., 2016] Nguyen, K., Sahin, O., Stewart, R., and Zhang, H. (2016). AUTOFLOW©- A novel application for water resource management and climate change response using smart technology. In Environmental Modelling and Software for Supporting a Sustainable Future, Proceedings of the 8th International Congress on Environmental Modelling and Software (iEMSs 2016). iEMSs. [Niknam et al., 2023a] Niknam, A., Zare, H. K., Hosseininasab, H., and Mostafaeipour, A. (2023a). A hybrid approach combining the multi-dimensional time series k-means algorithm and long short-term memory networks to predict the monthly water demand according to the uncertainty in the dataset. Earth Science Informatics, 16(2):1519–1536. [Niknam et al., 2023b] Niknam, A., Zare, H. K., Hosseininasab, H., and Mostafaeipour, A. (2023b). Developing an LSTM model to forecast the monthly water consumption according to the effects of the climatic factors in Yazd, Iran. Journal of Engineering Research, 11(1):100028. [Niyongabo et al., 2024] Niyongabo, A., Zhang, D., Guan, Y., Wang, Z., Imran, M., Nicayenzi, B., Guyasa, A. K., and Hatungimana, P. (2024). Predicting Urban Water Consumption and Health Using Artificial Intelligence Techniques in Tanganyika Lake, East Africa. Water, 16(13). [Oyebode and Ighravwe, 2019] Oyebode, O. and Ighravwe, D. E. (2019). Urban Water Demand Forecasting: A Comparative Evaluation of Conventional and Soft Computing Techniques. Resources, 8(3). [Pakpahan et al., 2023] Pakpahan, G., Rahayu, L., and Adhim, F. (2023). Prediction of Clean Water Needs in Surabaya Using the Time - Series Neural Network Method. In 2023 International Conference on Advanced Mechatronics, Intelligent Manufacture and Industrial Automation, ICAMIMIA 2023 - Proceedings, 2023 International Conference on Advanced Mechatronics, Intelligent Manufacture and Industrial Automation, ICAMIMIA 2023 - Proceedings, pages 656–661, United States. Institute of Electrical and Electronics Engineers Inc. [Polić et al., 2023] Polić, T., Mutka, A., and Žagar, M. (2023). Time Series LSTM Prediction of Water Consumption in the Republic of Croatia by Regions. In 2023 International Conference on Software, Telecommunications and Computer Networks (SoftCOM), pages 1–6. [Pourmousavi et al., 2022] Pourmousavi, M., Nasrollahi, H., Najafabadi, A. A., and Kalhor, A. (2022). Evaluating the performance of feature selection techniques and machine learning algorithms on future residential water demand. Water Supply, 22(8):6833–6854. 8 [Pu et al., 2022] Pu, Z., Yan, J., Chen, L., Li, Z., Tian, W., Tao, T., and Xin, K. (2022). A hybrid Wavelet-CNN-LSTM deep learning model for short-term urban water demand forecasting. Frontiers of Environmental Science & Engineering, 17(2):22. [Rustam et al., 2022] Rustam, F., Ishaq, A., Kokab, S. T., de la Torre Diez, I., Mazón, J. L. V., Rodríguez, C. L., and Ashraf, I. (2022). An Artificial Neural Network Model for Water Quality and Water Consumption Prediction. Water, 14(21). [Sahoo et al., 2023] Sahoo, B. B., Panigrahi, B., Nanda, T., Tiwari, M. K., and Sankalp, S. (2023). Multi-step Ahead Urban Water Demand Forecasting Using Deep Learning Models. SN Computer Science, 4(6):752. [Said et al., 2021] Said, N. M., Zin, Z. M., Ismail, M. N., and Bakar, T. A. (2021). Univariate water consumption time series prediction using deep learning in neural network (dlnn). International Journal of Advanced Technology and Engineering Exploration, 8(76):473–483. [Sajadifar and Pakseresht, 2024] Sajadifar, S. and Pakseresht, M. (2024). A Probabilistic Markov Chain Model for Shortterm Water Demand Forecasting. Journal of Water Management Modeling. [Sajjanshetty et al., 2023] Sajjanshetty, A. S., Jayanth, V., Mohan, R., Pahari, S., and Deepti, C. (2023). Estimation of Community Water Consumption Using Multivariate Ensemble Approach. In 2023 IEEE International Conference on Contemporary Computing and Communications (InC4), pages 1–5. [Santos de Jesus and Silva Gomes, 2023] Santos de Jesus, E. d. and Silva Gomes, G. S. d. (2023). Machine learning models for forecasting water demand for the Metropolitan Region of Salvador, Bahia. Neural Computing and Applications, 35(27):19669–19683. [Sebri, 2013] Sebri, M. (2013). ANN versus SARIMA models in forecasting residential water consumption in Tunisia. Journal of Water Sanitation and Hygiene for Development, 3:330–340. [Sebri, 2016] Sebri, M. (2016). Forecasting urban water demand: A meta-regression analysis. Journal of Environmental Management, 183:777–785. [Shi et al., 2013] Shi, P., Yang, T., Chen, X., Yu, Z., Acharyad, K., and Xu, C. (2013). Urban water consumption in a rapidly developing flagship megacity of South China: prospective scenarios and implications. Stochastic Environmental Research and Risk Assessment, 27(6):1359–1370. [Smolak et al., 2020] Smolak, K., Kasieczka, B., Fialkiewicz, W., Rohm, W., Siła-Nowicka, K., and and, K. K. (2020). Applying human mobility and water consumption data for short-term water demand forecasting using classical and machine learning models. Urban Water Journal, 17(1):32–42. [Suh and Ham, 2016] Suh, D. and Ham, S. (2016). A water demand forecasting model using BPNN for residential building. Contemporary Engineering Sciences, 9(1):1–10. [Thakur et al., 2021] Thakur, T., Mehra, A., Hassija, V., Chamola, V., Srinivas, R., Gupta, K. K., and Singh, A. P. (2021). Smart water conservation through a machine learning and blockchain-enabled decentralized edge computing network. Applied Soft Computing, 106:107274. [Uzlu, 2024] Uzlu, E. (2024). Modeling and Forecasting of Water Demand in the City of Istanbul Using Artificial Neural Networks Optimized with Rao Algorithms. Arabian Journal for Science and Engineering, 49(10):13477–13490. [Velasco et al., 2018] Velasco, L. C. P., Granados, A. R. B., Ortega, J. M. A., and Pagtalunan, K. V. D. (2018). Performance Analysis of Artificial Neural Networks Training Algorithms and Transfer Functions for Medium-Term Water Consumption Forecasting. International Journal of Advanced Computer Science and Applications, 9(4). [Vo et al., 2022] Vo, M. T., Vu, D., Nguyen, H., Bui, H., and Le, T. (2022). Predicting Monthly Household Water Consumption. In 2022 RIVF International Conference on Computing and Communication Technologies (RIVF), pages 726–730. [Wang et al., 2023] Wang, C., Li, Z., Ni, X., Shi, W., Zhang, J., Bian, J., and Liu, Y. (2023). Residential water and energy consumption prediction at hourly resolution based on a hybrid machine learning approach. Water Research, 246:120733. [Wang et al., 2024] Wang, Q., Wang, P., and Cai, M. (2024). Multivariate time series forecasting of daily urban water demand using reinforcement learning and gated recurrent unit network. In Proceedings of the 2024 7th International Conference on Data Storage and Data Engineering, DSDE ’24, page 45–51, New York, NY, USA. Association for Computing Machinery. [Xenochristou et al., 2021] Xenochristou, M., Hutton, C., Hofman, J., and Kapelan, Z. (2021). Short-Term Forecasting of Household Water Demand in the UK Using an Interpretable Machine Learning Approach. Journal of Water Resources Planning and Management, 147(4):04021004. 9