Journal of Intelligence and Information Systems (지능정보연구)

Korea Intelligent Information System Society (한국지능정보시스템학회)
권호 표지
DOI QR코드

DOI QR Code

Application of spatiotemporal transformer model to improve prediction performance of particulate matter concentration

미세먼지 예측 성능 개선을 위한 시공간 트랜스포머 모델의 적용

  • Kim, Youngkwang (Data Solution Business Department, WesleyQuest Co., Ltd.) ;
  • Kim, Bokju (D&A Platform Department, Woori Finance Information System Co., Ltd.) ;
  • Ahn, SungMahn (School of Business Administration, Kookmin University)
  • Received : 2022.03.03
  • Accepted : 2022.03.25
  • Published : 2022.03.31
Download PDF
⟨ Previous Next ⟩

Abstract

It is reported that particulate matter(PM) penetrates the lungs and blood vessels and causes various heart diseases and respiratory diseases such as lung cancer. The subway is a means of transportation used by an average of 10 million people a day, and although it is important to create a clean and comfortable environment, the level of particulate matter pollution is shown to be high. It is because the subways run through an underground tunnel and the particulate matter trapped in the tunnel moves to the underground station due to the train wind. The Ministry of Environment and the Seoul Metropolitan Government are making various efforts to reduce PM concentration by establishing measures to improve air quality at underground stations. The smart air quality management system is a system that manages air quality in advance by collecting air quality data, analyzing and predicting the PM concentration. The prediction model of the PM concentration is an important component of this system. Various studies on time series data prediction are being conducted, but in relation to the PM prediction in subway stations, it is limited to statistical or recurrent neural network-based deep learning model researches. Therefore, in this study, we propose four transformer-based models including spatiotemporal transformers. As a result of performing PM concentration prediction experiments in the waiting rooms of subway stations in Seoul, it was confirmed that the performance of the transformer-based models was superior to that of the existing ARIMA, LSTM, and Seq2Seq models. Among the transformer-based models, the performance of the spatiotemporal transformers was the best. The smart air quality management system operated through data-based prediction becomes more effective and energy efficient as the accuracy of PM prediction improves. The results of this study are expected to contribute to the efficient operation of the smart air quality management system.

미세먼지는 폐나 혈관에 침투해 각종 심장 질환이나 폐암 등의 호흡기 질환을 일으키는 것으로 보고되고 있다. 지하철은 일 평균 천만 명이 이용하는 교통수단으로, 깨끗하고 쾌적한 환경조성이 중요하나 지하터널을 통과하는 지하철의 운행 특성과 터널에 갇힌 미세먼지가 열차 풍으로 인해 지하역사로 이동하는 등의 문제로 지하역사의 미세먼지 오염도는 높은 것으로 나타나고 있다. 환경부와 서울시는 지하역사 공기질 개선대책을 수립하여 다양한 미세먼지 저감 노력을 기울이고 있다. 스마트 공기질 관리 시스템은 공기질 데이터 수집 및 미세먼지 농도를 예측하여 공기질을 관리하는 시스템으로 미세먼지 농도 예측 모델이 중요한 구성 요소이다. 그동안 시계열 데이터 예측에 관한 다양한 연구가 진행되어왔지만, 지하철 역사의 미세먼지 농도 예측과 관련해서는 통계나 순환신경망 기반의 딥러닝 모델 연구에 국한되어 있다. 이에 본 연구에서는 시공간 트랜스포머를 포함한 4개의 트랜스포머 기반 모델을 제안한다. 서울시 지하철 역사의 대합실을 대상으로 한 시간 후의 미세먼지 농도 예측실험을 수행한 결과, 트랜스포머 기반 모델들의 성능이 기존의 ARIMA, LSTM, Seq2Seq 모델들에 비해 우수한 성능을 나타냄을 확인하였다. 트랜스포머 기반 모델 중에서는 시공간 트랜스포머의 성능이 가장 우수하였다. 데이터 기반의 예측을 통하여 운영되는 스마트 공기질 관리 시스템은 미세먼지 예측의 정확도가 향상될수록 더욱더 효과적이고 에너지 효율적으로 운영될 수 있다. 본 연구 결과는 스마트 공기질 관리 시스템의 효율적 운영에 기여할 수 있을 것으로 기대된다.

Keywords

References

  1. 권순박. (2018). 공기청정기 실증 사례 분석-지하역사 스마트 공기질관리시스템 연구 사례. Air Cleaning Technology, 31(3), 39-46.
  2. 권순박, 강중구, 류승원, 남궁형규, 박세찬, 김민해, 김진호. (2017). 스마트 철도역사의 인공지능기반 실내공기질 관리기술. 한국철도학회 학술발표대회논문집, 578-580.
  3. 김인경, 김대희, 이재구. (2021). Temporal fusion transformer 모델을 활용한 다층 수평 시계열 데이터 분석. 한국정보처리학회 학술대회논문집, 28(1), 479-482.
  4. 서울특별시청. (2022년, 1월 27일). 지하역사 공기질 개선(2022), Retrieved 2월 26일, 2022년, from https://yesan.seoul.go.kr/wk/wkSelect.do?itemId=106477
  5. 안성만, 정여진, 이재준, 양지헌. (2017). 한국어음소 단위 lstm 언어모델을 이용한 문장 생성. 지능정보연구, 23(2), 71-88. https://doi.org/10.13088/JIIS.2017.23.2.071
  6. 오종민, 신현수, 신예슬, 정형철. (2017). 시계열 분석을 활용한 서울시 미세먼지 예측. Journal of the Korean Data Analysis Society (JKDAS), 19(5), 2457-2468. https://doi.org/10.37727/jkdas.2017.19.5.2457
  7. 이정영, 이종현, 이영재, 김록호, 한진석. (2007). ARIMA 모형을 이용한 서울지역 O3 오염도의 시계열 분석. 한국대기환경학회 학술대회논문집, 1363-1366.
  8. 이현욱. (2020). 지하철 실내 공기 질 개선을 위한 철도차량 휠-레일 접촉 미세마모입자 발생 연구 소개. 공업화학전망, 23(4), 20-29.
  9. 정철우, 김명석. (2013). Comparison studies of hybrid and non-hybrid forecasting models for seasonal and trend time series data. 지능정보연구, 19(1), 1-17. https://doi.org/10.13088/JIIS.2013.19.1.001
  10. 진세종, 조형준. (2020). 머신러닝을 활용한 계절 시계열 예측. Journal of the Korean Data Analysis Society, 22(5), 1779-1791. https://doi.org/10.37727/jkdas.2020.22.5.1779
  11. 질병관리청. (2020년, 3월 23일). 미세먼지. Retrieved 2월 15일, 2022년, from https://www.kdca.go.kr/contents.es?mid=a20304030300
  12. 차진욱, 김장영. (2018). 미세먼지 수치 예측 모델 구현을 위한 데이터마이닝 알고리즘 개발. 한국정보통신학회논문지, 22(4), 595-601. https://doi.org/10.6109/jkiice.2018.22.4.595
  13. 환경부. (2018년, 8월 31일). 제3차 지하역사 공기질 개선대책. Retrieved 2월 15일, 2022년, from http://www.me.go.kr/home/web/policy_data/read.do?menuId=10276&seq=7188
  14. 홍성원. (2020). 오토인코더 기반 특징추출을 통한 국가별 covid-19 일일 신규 확진자 수 예측 모델링. 한국지능정보시스템학회 학술대회논문집, 2020(6), 57-58.
  15. Bahdanau, D., Cho, K., & Bengio, Y. (2014). Neural machine translation by jointly learning to align and translate. arXiv Preprint arXiv: 1409.0473.
  16. Bengio, Y., Simard, P., & Frasconi, P. (1994). Learning long-term dependencies with gradient descent is difficult. IEEE Transactions on Neural Networks, 5(2), 157-166. https://doi.org/10.1109/72.279181
  17. Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time series analysis: Forecasting and control. John Wiley & Sons.
  18. Bui, T., Le, V., & Cha, S. (2018). A deep learning approach for forecasting air pollution in south korea using LSTM. arXiv Preprint arXiv: 1804.07891.
  19. Chen, K., Chen, G., Xu, D., Zhang, L., Huang, Y., & Knoll, A. (2021). NAST: Non-autoregressive spatial-temporal transformer for time series forecasting. arXiv Preprint arXiv:2102.05624.
  20. Chen, T., Yin, H., Chen, H., Wu, L., Wang, H., Zhou, X., & Li, X. (2018). Tada: Trend alignment with dual-attention multi-task recurrent neural networks for sales prediction. Paper presented at the 2018 IEEE International Conference on Data Mining (ICDM), 49-58.
  21. Cho, K., Van Merrienboer, B., Bahdanau, D., & Bengio, Y. (2014a). On the properties of neural machine translation: Encoder-decoder approaches. arXiv Preprint arXiv:1409.1259.
  22. Cho, K., Van Merrienboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014b). Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv Preprint arXiv:1406.1078.
  23. Goyal, P., Chan, A. T., & Jaiswal, N. (2006). Statistical models for the prediction of respirable suspended particulate matter in urban cities. Atmospheric Environment, 40(11), 2068-2077. https://doi.org/10.1016/j.atmosenv.2005.11.041
  24. Grigsby, J., Wang, Z., & Qi, Y. (2021). Long-range transformers for dynamic spatiotemporal forecasting. arXiv Preprint arXiv:2109.12218.
  25. Hamid, T. S., & Sodoudi, S. (2016). Statistical modeling approaches for PM10 prediction in urban areas; A review of 21st-century studies. Atmosphere, 7(2), 15.
  26. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735
  27. Hyndman, R., Koehler, A. B., Ord, J. K., & Snyder, R. D. (2008). Forecasting with exponential smoothing: The state space approach Springer. Science & Business Media.
  28. Kazemi, S. M., Goel, R., Eghbali, S., Ramanan, J., Sahota, J., Thakur, S., ... Brubaker, M. (2019). Time2Vec: Learning a vector representation of time. arXiv Preprint arXiv: 1907.05321
  29. Lara-Benitez, P., Carranza-Garcia, M., & Riquelme, J. C. (2021). An experimental review on deep learning architectures for time series forecasting. International Journal of Neural Systems, 31(3), undefined.
  30. Lee, S., Liu, H., Kim, M., Kim, J. T., & Yoo, C. (2014). Online monitoring and interpretation of periodic diurnal and seasonal variations of indoor air pollutants in a subway station using parallel factor analysis (PARAFAC). Energy and Buildings, 68, 87-98. https://doi.org/10.1016/j.enbuild.2013.09.022
  31. Li, S., Jin, X., Xuan, Y., Zhou, X., Chen, W., Wang, Y., & Yan, X. (2019). Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. arXiv Preprint arXiv:1907.00235.
  32. Lim, B., Arik, S. O., Loeff, N., & Pfister, T. (2019). Temporal fusion transformers for interpretable multi-horizon time series forecasting. International Journal of Forecasting, 37(4), 1748-1764.
  33. Liu, L., Liu, J., & Han, J. (2021). Multi-head or single-head? an empirical comparison for transformer training. arXiv Preprint arXiv: 2106.09650.
  34. Loy-Benitez, J., Vilela, P., Li, Q., & Yoo, C. (2019). Sequential prediction of quantitative health risk assessment for the fine particulate matter in an underground facility using deep recurrent neural networks. Ecotoxicology and Environmental Safety, 169, 316-324. https://doi.org/10.1016/j.ecoenv.2018.11.024
  35. Luong, M., Pham, H., & Manning, C. D. (2015). Effective approaches to attention-based neural machine translation. arXiv Preprint arXiv: 1508.04025.
  36. Martinez-Alvarez, F., Troncoso, A., Asencio-Cortes, G., & Riquelme, J. C. (2015). A survey on data mining techniques applied to electricity-related time series forecasting. Energies, 8(11), 13162-13193. https://doi.org/10.3390/en81112361
  37. Olah, C. (2015, August 27). Understanding LSTM Networks, Retrieved February 15, 2022, from http://colah.github.io/posts/2015-08-Understanding-LSTMs/
  38. Padhi, I., Schiff, Y., Melnyk, I., Rigotti, M., Mroueh, Y., Dognin, P., ... Altman, E. (2021). Tabular transformers for modeling multivariate time series. Paper presented at the ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 3565-3569.
  39. Park, S., Kim, M., Kim, M., Namgung, H., Kim, K., Cho, K. H., & Kwon, S. (2018). Predicting PM10 concentration in seoul metropolitan subway stations using artificial neural network (ANN). Journal of Hazardous Materials, 341, 75-82. https://doi.org/10.1016/j.jhazmat.2017.07.050
  40. Quintana Valenzuela, D. (2021). A study of deep learning techniques for sequence-based problems. Master in Research in Informatics, Universitat Politecnica de Catalunya, Barcelona.
  41. Reddy, V., Yedavalli, P., Mohanty, S., & Nakhat, U. (2018). Deep air: Forecasting air pollution in beijing, china. Environmental Science, Retrieved 1 November, 2021, from https://www.ischool.berkeley.edu/sites/default/files/sproject_attachments/deep-air-forecasting_final.pdf
  42. Soh, P., Chang, J., & Huang, J. (2018). Adaptive deep learning-based air quality prediction model using the most relevant spatial-temporal relations. IEEE Access, 6, 38186-38199. https://doi.org/10.1109/access.2018.2849820
  43. Smith, T. G. (2017). Pmdarima: ARIMA estimators for Python. 3 January, 2022, from http://www.alkaline-ml.com/pmdarima
  44. Suilin, A. (2018, May 9). How it works. Retrieved February 26, 2022, from https://github.com/Arturus/kaggle-web-traffic/blob/master/how_it_works.md
  45. Sutskever, I., Vinyals, O., & Le, Q. V. (2014). Sequence to sequence learning with neural networks. Advances in Neural Information Processing Systems, 27
  46. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., ... Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30
  47. Wass, D. (2021). Transformer learning for traffic prediction in mobile networks. Degree Project in Computer Science and Engineering, KTH Royal Institute of Technology, Stockholm.
  48. Wu, N., Green, B., Ben, X., & O'Banion, S. (2020). Deep transformer models for time series forecasting: The influenza prevalence case. arXiv Preprint arXiv:2001.08317.