Journal of Korea Water Resources Association (한국수자원학회논문집)

Korea Water Resources Association (한국수자원학회)
권호 표지
DOI QR코드

DOI QR Code

Performance improvement of artificial neural network based water quality prediction model using explainable artificial intelligence technology

설명가능한 인공지능 기술을 이용한 인공신경망 기반 수질예측 모델의 성능향상

  • Lee, Won Jin (Department of Civil Engineering, Chungbuk National University) ;
  • Lee, Eui Hoon (Department of Civil Engineering, Chungbuk National University)
  • 이원진 (충북대학교 토목공학과) ;
  • 이의훈 (충북대학교 토목공학부)
  • Received : 2023.10.04
  • Accepted : 2023.11.07
  • Published : 2023.11.30
Download PDF
⟨ Previous Next ⟩

Abstract

Recently, as studies about Artificial Neural Network (ANN) are actively progressing, studies for predicting water quality of rivers using ANN are being conducted. However, it is difficult to analyze the operation process inside ANN, because ANN is form of Black-box. Although eXplainable Artificial Intelligence (XAI) is used to analyze the computational process of ANN, research using XAI technology in the field of water resources is insufficient. This study analyzed Multi Layer Perceptron (MLP) to predict Water Temperature (WT), Dissolved Oxygen (DO), hydrogen ion concentration (pH) and Chlorophyll-a (Chl-a) at the Dasan water quality observatory in the Nakdong river using Layer-wise Relevance Propagation (LRP) among XAI technologies. The MLP that learned water quality was analyzed using LRP to select the optimal input data to predict water quality, and the prediction results of the MLP learned using the optimal input data were analyzed. As a result of selecting the optimal input data using LRP, the prediction accuracy of MLP, which learned the input data except daily precipitation in the surrounding area, was the highest. Looking at the analysis of MLP's DO prediction results, it was analyzed that the pH and DO a had large influence at the highest point, and the effect of WT was large at the lowest point.

최근 인공신경망(Artificial Neural Network, ANN)의 연구가 활발하게 진행되면서 ANN을 이용하여 하천의 수질을 예측하는 연구가 진행되고 있다. 그러나 ANN은 Black-box의 형태이기 때문에 ANN 내부의 연산과정을 분석하는데 어려움이 있다. ANN의 연산과정을 분석하기 위해 설명가능한 인공지능(eXplainable Artificial Intelligence, XAI) 기술이 사용되고 있으나, 수자원 분야에서 XAI 기술을 활용한 연구는 미비한 실정이다. 본 연구는 XAI 기술 중 Layer-wise Relevance Propagation (LRP)을 사용하여 낙동강의 다산 수질관측소의 수온, 용존산소량, 수소이온농도 및 엽록소-a를 예측하기 위한 Multi Layer Perceptron (MLP)을 분석하였다. LRP를 기반으로 수질을 학습한 MLP를 분석하여 수질을 예측하기 위한 최적의 입력자료를 선정하고, 최적의 입력자료를 이용하여 학습한 MLP의 예측결과에 대한 분석을 실시하였다. LRP를 이용하여 최적의 입력자료를 선정한 결과를 보면, 수온, 용존산소량, 수소이온농도 및 엽록소-a 모두 주변지역의 일 강수량을 제외한 입력자료를 학습한 MLP의 예측정확도가 가장 높았다. MLP의 용존산소량 예측결과에 대한 분석결과를 보면, 최고점에서 수소이온농도 및 용존산소량의 영향이 크고 최저점에서는 수온의 영향이 큰 것으로 분석되었다.

Keywords

Acknowledgement

이 논문은 충북대학교 국립대학육성사업(2023)지원을 받아 작성되었음

References

  1. Ahn, S.J., Yeon, I.S., Han, Y.S., and Lee, J.K. (2001). "Water quality forecasting at Gongju station in Geum River using neural network model." Journal of Korea Water Resources Association, Vol. 34, No. 6, pp. 701-711. 
  2. Ahn, S.R., Ha, R., Yoon, S.W., and Kim, S.J. (2014). "Evaluation of future turbidity water and eutrophication in Chungju Lake by climate change using CE-QUAL-W2." Journal of Korea Water Resources Association, Vol. 47, No. 2, pp. 145-159.  https://doi.org/10.3741/JKWRA.2014.47.2.145
  3. Arras, L., Montavon, G., Muller, K.R., and Samek, W. (2017). "Explaining recurrent neural network predictions in sentiment analysis." arXiv, arXiv:1706.07206. 
  4. Bach, S., Binder, A., Montavon, G., Klauschen, F., Muller, K.R., and Samek, W. (2015). "On pixel-wise explanations for nonlinear classifier decisions by layer-wise relevance propagation." PloS One, Vol. 10, No. 7, e0130140. 
  5. Carbone, G., Bortolussi, L., and Sanguinetti, G. (2022). "Resilience of bayesian layer-wise explanations under adversarial attacks." Proceedings of the 2022 International Joint Conference on Neural Networks (IJCNN), Padua, Italy, pp. 1-8. 
  6. Chun, Y.E., Kim, S.B., Lee, J.Y., and Woo, J.H. (2021). "Study on credit rating model using explainable AI." The Korean Data & Information Science Society, Vol. 32, No. 2, pp. 283-295.  https://doi.org/10.7465/jkdi.2021.32.2.283
  7. Chung, J., Gulcehre, C., Cho, K., and Bengio, Y. (2014). "Empirical evaluation of gated recurrent neural networks on sequence modeling." arXiv, arXiv:1412.3555. 
  8. Dogan, E., Ates, A., Yilmaz, E.C., and Eren, B. (2008). "Application of artificial neural networks to estimate wastewater treatment plant inlet biochemical oxygen demand." Environmental Progress, Vol. 27, No. 4, pp. 439-446.  https://doi.org/10.1002/ep.10295
  9. Drolc, A., and Koncan, J.Z. (1999). "Calibration of QUAL2E model for the Sava River (Slovenia)." Water Science and Technology, Vol. 40, No. 10, pp. 111-118.  https://doi.org/10.2166/wst.1999.0509
  10. Haghiabi, A.H., Nasrolahi, A.H., and Parsaie, A. (2018). "Water quality prediction using machine learning methods." Water Quality Research Journal, Vol. 53, No. 1, pp. 3-13.  https://doi.org/10.2166/wqrj.2018.025
  11. Hochreiter, S., and Schmidhuber, J. (1997). "Long short-term memory." Neural Computation, Vol. 9, No. 8, pp. 1735-1780.  https://doi.org/10.1162/neco.1997.9.8.1735
  12. Kim, S.H., Park, J.H., and Kim, B. (2021). "Prediction of cyanobacteria harmful algal blooms in reservoir using machine learning and deep learning." Journal of Korea Water Resources Association, Vol. 54, No. 12, pp. 1167-1181.  https://doi.org/10.3741/JKWRA.2021.54.S-1.1167
  13. Lee, J., and Han, J. (2021). "Layer-wise Relevance Propagation (LRP) based technical and macroeconomic indicator impact analysis for an explainable deep learning model to predict an increase and decrease in KOSPI." Journal of Korean Institute of Information Scientists and Engineers, Vol. 48, No. 12, pp. 1289-1297.  https://doi.org/10.5626/JOK.2021.48.12.1289
  14. Lee, J.H., Kim, J.S., Jang, H.W., and Lee, J.C. (2013). "Drought forecasting using the multi layer perceptron (MLP) artificial neural network model." Journal of Korea Water Resources Association, Vol. 46, No. 12, pp. 1249-1263.  https://doi.org/10.3741/JKWRA.2013.46.12.1249
  15. Lee, S., and Lee, D. (2018). "Improved prediction of harmful algal blooms in four Major South Korea's Rivers using deep learning models." International Journal of Environmental Research and Public Health, Vol. 15, No. 7, 1322. 
  16. Lee, W.J., and Lee, E.H. (2023). "Improvement of multi layer perceptron performance using combination of adaptive moments and improved harmony search for prediction of Daecheong Dam inflow." Journal of Korea Water Resources Association, Vol. 56, No. 1, pp. 63-74.  https://doi.org/10.3741/JKWRA.2023.56.1.63
  17. Lim, H., An, H., Choi, E., and Kim, Y. (2020). "Prediction of the DO concentration using the machine learning algorithm: case study in Oncheoncheon, Republic of Korea." Korean Journal of Agricultural Science, Vol. 47, No. 4, pp. 1029-1037. 
  18. Lu, H., and Ma, X. (2020). "Hybrid decision tree-based machine learning models for short-term water quality prediction." Chemosphere, Vol. 249, 126169. 
  19. Luis, M.B., Sidinei, M.T., and Priscilla, C. (2010). "Limnological effects of Egeria najas Planchon (Hydrocharita-ceae) in the arms of Itaipu Reservoir (Brazil, Paraguay)." Limnology, Vol. 11, No. 1, pp. 39-47.  https://doi.org/10.1007/s10201-009-0286-4
  20. Lundberg, S.M., and Lee, S.I. (2017). "A unified approach to interpreting model predictions." Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, U.S., Vol. 30. pp. 4768-4777. 
  21. Luo, D.L. (2002). "Study on the distribution of dissolved oxygen in Shenhu Bay and its relationship with phytoplankton and suspended matter." Marine Science Bulletin, Vol. 21, No. 1, pp. 31-36. 
  22. Mahsa, M., and Lee, T. (2018). "Comparison of optimization algorithms in deep learning-based neural networks for hydrological forecasting: Case study of Nam River daily runoff." Journal of the Korean Society of Hazard Mitigation, Vol. 18, No. 6, pp. 377-384.  https://doi.org/10.9798/KOSHAM.2018.18.6.377
  23. McCulloch, W.S., and Pitts, W. (1943). "A logical calculus of the ideas immanent in nervous activity." The Bulletin of Mathematical Biophysics, Vol. 5, pp. 115-133.  https://doi.org/10.1007/BF02478259
  24. Mok, J.Y., Choi, J.H., and Moon, Y.I. (2020). "Prediction of multipurpose dam inflow using deep learning." Journal of Korea Water Resources Association, Vol. 53, No. 2, pp. 97-105. 
  25. Nahm, E-S. (2022). "Neural network modeling based XAI of activated sludge process in wastewater treatment system for dissolved oxygen control." The Transactions of the Korean Institute of Electrical Engineers, Vol. 71, No. 8, pp. 1176-1181.  https://doi.org/10.5370/KIEE.2022.71.8.1176
  26. Nawi, N.M., Atomi, W.H., and Rehman, M.Z. (2013). "The effect of data pre-processing on optimized training of artificial neural networks." Procedia Technology, Vol. 11, pp. 32-39.  https://doi.org/10.1016/j.protcy.2013.12.159
  27. Park, S.Y., Choi, J.H., Wang, S., and Park, S.S. (2006). "Design of a water quality monitoring network in a large river system using the genetic algorithm." Ecological Modelling, Vol. 199, No. 3, pp. 289-297.  https://doi.org/10.1016/j.ecolmodel.2006.06.002
  28. Ribeiro, M.T., Singh, S., and Guestrin, C. (2016). ""Why should i trust you?" Explaining the predictions of any classifier." Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, U.S., pp. 1135-1144. 
  29. Roh, S., and Park, D. (2021). "Sweet persimmons classification based on a mixed two-step synthetic neural network." Journal of Korea Multimedia Society, Vol. 24, No. 10, pp. 1358-1368. 
  30. Rosenblatt, F. (1958). "The perceptron: A probabilistic model for information storage and organization in the brain." Psychological review, Vol. 65, No. 6, pp. 386-408. https://doi.org/10.1037/h0042519
  31. Rumelhart, D.E., Hinton, G.E., and Williams, R.J. (1986). "Learning representations by back-propagating errors." Nature, Vol. 323, No. 6088, pp. 533-536.  https://doi.org/10.1038/323533a0
  32. Sheng, T.Q., and Xu, Y.Z. (1993). "Distribution of dissolved oxygen and pH in Kuroshio area of East of China Sea." Marine Science Bulletin, Vol. 12, No. 4, pp. 55-62. 
  33. Wang, T.S., Tan, C.H., Chen, L., and Tsai, Y.C. (2008). "Applying artificial neural networks and remote sensing to estimate chlorophyll-a concentration in water body." Proceedings 2008 Second International Symposium on Intelligent Information Technology Application, IEEE, Shanghai, China, Vol. 1, pp. 540-544. 
  34. Wu, H., Huang, A., and Sutherland, J.W. (2022). "Layer-wise relevance propagation for interpreting LSTM-RNN decisions in predictive maintenance." The International Journal of Advanced Manufacturing Technology, Vol. 118, pp. 963-978.  https://doi.org/10.1007/s00170-021-07911-9
  35. Yoo, Y., Kim, D., and Lee, J. (2020). "Performance analysis of various activation functions in super resolution model." Proceedings of the Korea Information Processing Society Conference, Vol. 27, No. 1, pp. 504-507. 
  36. Zhou, T., Jiang, Z., Liu, X., and Tan, K. (2020). "Research on the long-term and short-term forecasts of navigable river's water-level fluctuation based on the adaptive multilayer perceptron." Journal of Hydrology, Vol. 591, 125285.