Machine learning approaches for non-intrusive load monitoring: from qualitative to quantitative comparation


Non-intrusive load monitoring (NILM) is the prevailing method used to monitor the energy profile of a domestic building and disaggregate the total power consumption into consumption signals by appliance. Whilst the most popular disaggregation algorithms are based on Hidden Markov Model solutions based on deep neural networks have attracted interest from researchers. The objective of this paper is to provide a comprehensive overview of the NILM method and present a comparative review of modern approaches. In this effort, many obstacles are identified. The plethora of metrics, the variety of datasets and the diversity of methodologies make an objective comparison almost impossible. An extensive analysis is made in order to scrutinize these problems. Possible solutions and improvements are suggested, while future research directions are discussed.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2


  1. 1.

    The authors would like to thank Odyssefs Krystalakos for his contribution in developing the software and running the experiments for the appliance set complexity. The source code is available at:


  1. Aiad M, Lee PH (2016) Unsupervised approach for load disaggregation with devices interactions. Energy Build 116:96–103.

    Article  Google Scholar 

  2. Anderson K, Ocneanu AF, Benitez D, Carlson D, Rowe A, Bergés M (2012) BLUED? A fully labeled public dataset for event-based non-intrusive load monitoring research. In: Proceedings of the 2nd KDD workshop on data mining applications in sustainability (SustKDD), Oct 2011, pp 1–5

  3. Armel CK, Gupta A, Shrimali G, Albert A (2013) Is disaggregation the holy grail of energy efficiency? The case of electricity. Energy Policy 52:213–234.

    Article  Google Scholar 

  4. Baranski M, Voss J (2003) Non-intrusive appliance load monitoring based on an optical sensor. In: 2003 IEEE Bologna power tech conference proceedings, vol 4, pp 267–274.

  5. Batra N, Kelly J, Parson O, Dutta H, Knottenbelt W, Rogers A, Srivastava M (2014) NILMTK? An open source toolkit for non-intrusive load monitoring categories and subject descriptors. In: International conference on future energy systems (ACM E-Energy), pp 1–4.

  6. Batra N, Parson O, Berges M, Singh A, Rogers A (2014) A comparison of non-intrusive load monitoring methods for commercial and residential buildings. Retrieved from arXiv:1408.6595 [Cs]

  7. Beckel C, Kleiminger W, Staake T, Santini S (2014) The ECO data set and the performance of non-intrusive load monitoring algorithms. In: Proceedings of the 1st ACM conference on embedded systems for energy-efficient buildings, pp 80–89.

  8. Bonfigli R, Squartini S, Fagiani M, Piazza F (2015) Unsupervised algorithms for non-intrusive load monitoring: an up-to-date overview. In: 2015 IEEE 15th international conference on environment and electrical engineering, EEEIC 2015—conference proceedings, pp 1175–1180.

  9. Burbano D (2015). Intrusive and non-intrusive load monitoring (a survey). Latin Am J Comput LAJC 2(1):45–53. Retrieved from

  10. Chang HH, Chien PC, Lin LS, Chen N (2011) Feature extraction of non-intrusive load-monitoring system using genetic algorithm in smart meters. In: Proceedings—2011 8th IEEE international conference on e-business engineering, ICEBE 2011, pp 299–304.

  11. Department of Energy and Climate Change (2008) Climate change act, Tech rep, UK Retrieved from

  12. Darby S (2006) The effectiveness of feedback on energy consumption a review for defra of the literature on metering, billing and direct displays. Retrieved from

  13. Dong R, Ratliff L, Ohlsson H, Sastry SS (2013) Fundamental limits of nonintrusive load monitoring. In: 3rd ACM international conference on high confidence networked systems (HiCoNS), pp 11–18.

  14. Egarter D, Bhuvana VP, Elmenreich W (2015) PALDi: online load disaggregation via particle filtering. IEEE Trans Instrum Meas 64(2):467–477.

    Article  Google Scholar 

  15. Egarter D, Pöchacker M, Elmenreich W (2015) Complexity of power draws for load disaggregation 1–26.

  16. Eia (2017) Annual Energy Outlook 2017 with projections to 2050. Retrieved from

  17. Froehlich J, Larson E, Gupta S, Cohn G, Reynolds M, Patel S (2011) Disaggregated end-use energy sensing for the smart grid. IEEE Perv Comput 10(1):28–39.

    Article  Google Scholar 

  18. Greveler U, Justus B, Loehr D (2012) Forensic content detection through power consumption. In: IEEE international conference on communications, pp 6759–6763.

  19. Hart GW (1992) Nonintrusive appliance load monitoring. Proc IEEE 80(12):1870–1891

    Article  Google Scholar 

  20. Inagaki S, Egami T, Suzuki T, Nakamura H, Ito K (2011) Nonintrusive appliance load monitoring based on integer programming. Electr Eng Jpn (English Translation of Denki Gakkai Ronbunshi) 174(2):1386–1392.

    Google Scholar 

  21. Johnson MJ, Willsky AS (2013) Bayesian nonparametric hidden semi-Markov models. 14:673–701. arXiv Preprint Retrieved fromarxiv:1203.1365

  22. Kamat PV (2007) Meeting the clean energy demand: nanostructure architectures for solar energy conversion.

  23. Kato T, Cho HS, Lee D, Toyomura T, Yamazaki T (2009) Appliance recognition from electric current signals for information-energy integrated network in home environments. In: Lecture notes in computer science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol 5597, LNCS, pp 150–157.

  24. Katos V, Vrakas D, Katsaros P (2011) A framework for access control with inference constraints. In: Proceedings—international computer software and applications conference, pp 289–297.

  25. Kelly J, Knottenbelt W (2015a) Neural NILM: deep neural networks applied to energy disaggregation. In: Proceedings of the 2nd ACM international conference on embedded systems for energy-efficient built environments. ACM, pp 55–64

  26. Kelly J, Knottenbelt W (2015b) The UK-DALE dataset, domestic appliance-level electricity demand and whole-house demand from five UK homes. Sci Data 2:150007.

    Article  Google Scholar 

  27. Kim H, Marwah M, Arlitt MF, Lyon G, Han J (2011) Unsupervised disaggregation of low frequency power measurements. In: Proceedings of the 11th SIAM international conference on data mining, pp 747–758.

  28. Kolter JZ, Johnson MJ (2011) REDD? A public data set for energy disaggregation research. SustKDD Workshop, no (1), pp 1–6. Retrieved from

  29. Kolter Z, Jaakkola T (2012) Approximate inference in additive factorial HMMs with application to energy disaggregation. In: Proceedings of the international conference on artificial intelligence and statistics, vol XX, pp 1472–1482. Retrieved from

  30. Krizhevsky A, Sutskever I, Geoffrey EH (2012) ImageNet classification with deep convolutional neural networks. In: Advances in neural information processing systems 25 (NIPS2012), pp 1–9.

  31. Lai YX, Lai CF, Huang YM, Chao HC (2013) Multi-appliance recognition system with hybrid SVM/GMM classifier in ubiquitous smart home. Inf Sci 230:39–55.

    Article  Google Scholar 

  32. Lange H, Bergés M (2016) The neural energy decoder: energy disaggregation by combining binary subcomponents. In: NILM2016 3rd international workshop on non-intrusive load monitoring. Retrieved from

  33. Laughman C, Lee K, Cox R, Shaw S, Leeb S, Norford L, Armstrong P (2003) Power signature analysis. IEEE Power Energy Mag 1(2):56–63.

    Article  Google Scholar 

  34. Liang J, Ng SKK, Kendall G, Cheng JWM (2010) Load signature studypart I: basic concept, structure, and methodology. IEEE Trans Power Deliv 25(2):551–560.

    Article  Google Scholar 

  35. Lin GY, Lee SC, Hsu JYJ, Jih WR (2010) Applying power meters for appliance recognition on the electric panel. In: Proceedings of the 2010 5th IEEE conference on industrial electronics and applications, ICIEA 2010, pp 2254–2259.

  36. Makonin S, Popowich F (2015) Nonintrusive load monitoring (NILM) performance evaluation: a unified approach for accuracy reporting. Energy Effic 8(4):809–814.

    Article  Google Scholar 

  37. Makonin S, Popowich F, Bajic IV, Gill B, Bartram L (2015) Exploiting HMM sparsity to perform online real-time nonintrusive load monitoring. IEEE Trans Smart Grid.

    Google Scholar 

  38. Makonin S, Popowich F, Bartram L, Gill B, Bajic IV (2013) AMPds: a public dataset for load disaggregation and eco-feedback research. In: 2013 IEEE electrical power and energy conference, EPEC 2013.

  39. Marchiori A, Hakkarinen D, Han Q, Earle L (2011) Circuit-level load monitoring for household energy management. IEEE Perv Comput 10(1):40–48.

    Article  Google Scholar 

  40. Mauch L, Yang B (2015) A new approach for supervised power disaggregation by using a deep recurrent LSTM network. In proceedings of the 3 rd IEEE global conference on signal and information processing (GlobalSIP), pp 63–67

  41. Paradiso F, Paganelli F, Giuli D, Capobianco S (2016) Context-based energy disaggregation in smart homes. Future Internet 8(1):4. Retrieved from

  42. Parson O, Ghosh S, Weal M, Rogers A (2011) Using hidden Markov models for iterative non-intrusive appliance monitoring. Electronics and Computer Science, University of Southampton, Hampshire, UK, pp 1–4. Retrieved from

  43. Parson O, Ghosh S, Weal M, Rogers A (2012) Non-intrusive load monitoring using prior models of general appliance types. In: Proceedings of the 26th AAAI conference on artificial intelligence, pp 356–362

  44. Parson O, Ghosh S, Weal M, Rogers A (2014) An unsupervised training method for non-intrusive appliance load monitoring. Artif Intell 217:1–19.

    Article  Google Scholar 

  45. Polat H, Du W, Renckes S, Oysal Y, Polat H, Renckes S, Du W (2010) Private predictions on hidden Markov models. Artif Intell Rev 34:53–72.

    Article  Google Scholar 

  46. Ruzzelli AG, Nicolas C, Schoofs A, O’Hare GMP (2010) Real-time recognition and profiling of appliances through a single electricity sensor. In: SECON 2010—2010 7th Annual IEEE communications society conference on sensor, mesh and ad hoc communications and networks.

  47. Sankar L, Raj Rajagopalan S, Mohajer S, Vincent Poor H (2013) Smart meter privacy: a theoretical framework. IEEE Trans Smart Grid 4(2):837–846.

    Article  Google Scholar 

  48. Srinivasan D, Ng WS, Liew AC (2006) Neural-network-based signature recognition for harmonic source identification. IEEE Trans Power Deliv 21(1):398–405

    Article  Google Scholar 

  49. Wytock M, Kolter JZ (2013) Contextually supervised source separation with application to energy disaggregation. In: Twenty-eighth AAAI conference on artificial intelligence, pp 1–10. Retrieved from arXiv:1312.5023

  50. Zachary CL (2015) A critical review of recurrent neural networks for sequence learning. arXiv Preprint, pp 1–35.

  51. Zeifman M (2012) Disaggregation of home energy display data using probabilistic approach. IEEE Trans Consum Electron 58(1):23–31.

    MathSciNet  Article  Google Scholar 

  52. Zeifman M, Roth K (2011) Nonintrusive appliance load monitoring: review and outlook. IEEE Trans Consum Electron 57(1):76–84.

    Article  Google Scholar 

  53. Zhang C, Zhong M, Wang Z, Goddard N, Sutton C (2017) Sequence-to-point learning with neural networks for non-intrusive load monitoring. Retrieved from

  54. Zhong M, Goddard N, Sutton C (2015) Latent Bayesian melding for integrating individual and population models. In: Advances in neural information processing systems, pp 3618–3626.

  55. Zia T, Bruckner D, Zaidi A (2011) A hidden Markov model based procedure for identifying household electric loads. In: IECON proceedings (industrial electronics conference), pp 3218–3223.

  56. Zoha A, Gluhak A, Imran MA, Rajasegarar S (2012) Non-intrusive load monitoring approaches for disaggregated energy sensing: a survey. Sensors (Switzerland) 12(12):16838–16866.

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Christoforos Nalmpantis.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Nalmpantis, C., Vrakas, D. Machine learning approaches for non-intrusive load monitoring: from qualitative to quantitative comparation. Artif Intell Rev 52, 217–243 (2019).

Download citation


  • Non-intrusive load monitoring (NILM)
  • Power disaggregation algorithms
  • Hidden Markov Model
  • Deep learning