簡易檢索 / 詳目顯示
| 研究生： |
江培文 Pei-Wen Chiang |
|---|---|
| 論文名稱： |
混合式時間效率預測模型應用於廣距時間序列資料 Hybrid Time-efficient Prediction Models for Wide Range Time-series Data |
| 指導教授： |
洪西進
Shi-Jinn Horng |
| 口試委員: |
范欽雄
吳怡樂 李正吉 楊朝棟 楊竹星 楊昌彪 洪西進 Shi-Jinn Horng |
| 學位類別： |
博士 Doctor |
| 系所名稱： |
電資學院 - 資訊工程系 Department of Computer Science and Information Engineering |
| 論文出版年： | 2021 |
| 畢業學年度： | 109 |
| 語文別： | 中文 |
| 論文頁數： | 74 |
| 中文關鍵詞： | 空洞卷積 、卷積神經網路 、空氣品質預測 、長短期記憶網路 、自動編碼器 、閘門遞迴單元 、細懸浮微粒預測 |
| 外文關鍵詞： | Dilated Convolution, Convolutional Neural Network, Air Quality Prediction, Long Short-term Memory Network, Autoencoder, Gated Recurrent Unit, PM2.5 Forecasting |
| 相關次數： | 點閱：357 下載：3 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
細懸浮微粒預測是全球重視且與人類健康高度相關的議題，精確且即時的空氣品質預測可以警示民眾留在室內且避免接近空氣污染的區域。本研究首先提出一個具時間效率的空洞卷積神經網路架構。本架構是由空洞卷積神經網路及長短期記憶網路所組成。本研究所採用的城市空氣品質資料集來自微軟研究院的城市空氣計畫，資料是由大陸北京36個監測站台所收集。該資料集包含二氧化氮、細懸浮微粒、二氧化硫、一氧化碳、臭氧、溫度、濕度、天氣、氣壓、風向和風速。實驗結果指出，相較於其他研究已提出的方法，本研究所提出的架構，可以顯著的減少模型訓練時間和預測誤差。
許多研究結果顯示出細懸浮微粒會導致人類的呼吸系統與心血管系統受損。因此，精確的預測細懸浮微粒不但可以提醒民眾遠離空氣品質惡化的區域，也可提供政府制定環境相關政策。本研究另外提出一個針對每日細懸浮微粒的混合式時間序列預測架構。本架構是由自動編碼器、空洞卷積神經網路及閘門遞迴單元所組成。本實驗的資料集取自環保署，由台灣76個監測站台所收集並使用於比較不同模型的預測結果。本研究所提出的方法，不但可運用於縣市區域範圍，亦可適用於單一特定站台以預測細懸浮微粒。本方法同時考量空氣品質、天候氣象與地理位置相關資料，因此可提升預測細懸浮微粒的精確度。此外，本細懸浮微粒預測模型亦能同時學習空間地理和時間趨勢的資料特徵。實驗結果顯示，本模型架構的預測精確度明顯優於其他預測方法。
PM2.5 concentration is a highly related to human health in the world. The accuracy of air quality prediction in time can warn the people to stay at indoor places and to stay away from the concentrated areas. In this work, a time-efficient dilated convolution framework is proposed. The proposed framework consists of the dilated convolutional neural networks, and the long short-term memory networks. The Urban Air Quality Dataset used in our work is from the Urban Air project of Microsoft Research, and there are thirty six monitoring stations in Beijing, China. The dataset contains NO2, PM2.5, SO2, CO, O3, temperature, humidity, weather, pressure, wind direction, and wind speed. Compared with the existing approaches, the experimental results demonstrate that the training time and prediction error are enormously decreased.
Many studies have shown that fine particulate matter harms the respiratory system and the cardiovascular system. Accurate PM2.5 forecasting can alert people to stay away from the concentrated areas and to provide the government with the environmental policies in the future. In this work, we also propose a hybrid time-series prediction framework for daily-based PM2.5 forecasting. The proposed framework consists of the autoencoders, the dilated convolutional neural networks, and the gated recurrent units. The experimental dataset with 76 monitoring stations from the Taiwan Environmental Protection Administration is applied to compare the baseline and the proposed models. The proposed model is applicable for the specified city-/county-wide region and also for the particular monitoring station/site to predict PM2.5 concentration. By considering air quality data, meteorological data, and geographical data simultaneously, the proposed model can increase the accuracy of PM2.5 prediction. In addition, the proposed PM2.5 forecasting model can learn the location-centric spatial features and the daily-based temporal features simultaneously. The experimental results show that the prediction accuracy of the proposed model is superior to those of the baseline models.
[1] Y. Zhang, M. Bocquet, V. Mallet, C. Seigneur, and A. Baklanov, "Real-time air quality forecasting, part I: History, techniques, and current status," Atmospheric Environ., vol. 60, pp. 632–655, Dec. 2012, doi: 10.1016/j.atmosenv.2012.06.031.
[2] Y. Zhang, M. Bocquet, V. Mallet, C. Seigneur, and A. Baklanov, "Real-time air quality forecasting, part II: State of the science, current research needs, and future prospects," Atmospheric Environ., vol. 60, pp. 656–676, Dec. 2012, doi: 10.1016/j.atmosenv.2012.02.041.
[3] S. Vardoulakis, B. E. A. Fisher, K. Pericleous, and N. Gonzalez-Flesca, "Modelling air quality in street canyons: A review," Atmospheric Environ., vol. 37, no. 2, pp. 155–182, Jan. 2003, doi: 10.1016/S1352-2310(02)00857-9.
[4] M. Dong, D. Yang, Y. Kuang, D. He, S. Erdal, and D. Kenski, "PM2.5 concentration prediction using hidden semi-Markov model-based times series data mining," Expert Syst. Appl., vol. 36, no. 5, pp. 9046–9055, Jul. 2009, doi: 10.1016/j.eswa.2008.12.017.
[5] A. Donnelly, B. Misstear, and B. Broderick, "Real time air quality forecasting using integrated parametric and non-parametric regression techniques," Atmospheric Environ., vol. 103, pp. 53–65, Feb. 2015, doi: 10.1016/j.atmosenv.2014.12.011.
[6] S. Duan, W. Song, E. Zio, C. Cattani, and M. Li, "Product technical life prediction based on multi-modes and fractional Lévy stable motion," Mech. Syst. Signal Process., vol. 161, p. 107974, Dec. 2021, doi: 10.1016/j.ymssp.2021.107974.
[7] H. Liu, W. Song, Y. Niu, and E. Zio, "A generalized cauchy method for remaining useful life prediction of wind turbine gearboxes," Mech. Syst. Signal Process., vol. 153, p. 107471, May 2021, doi: 10.1016/j.ymssp.2020.107471.
[8] H. Liu, W. Song, and E. Zio, "Metabolism and difference iterative forecasting model based on long-range dependent and grey for gearbox reliability," ISA Trans., May 2021, doi: 10.1016/j.isatra.2021.05.002.
[9] H. Liu, W. Song, Y. Zhang, and A. Kudreyko, "Generalized Cauchy degradation model with long-range dependence and maximum lyapunov exponent for remaining useful life," IEEE Trans. Instrum. Meas., vol. 70, pp. 1–12, Mar. 2021, doi: 10.1109/TIM.2021.3063749.
[10] H. Liu, W. Song, and E. Zio, "Generalized Cauchy difference iterative forecasting model for wind speed based on fractal time series," Nonlinear Dyn., vol. 103, no. 1, pp. 759–773, Jan. 2021, doi: 10.1007/s11071-020-06150-z.
[11] Y. Zheng, F. Liu, and H.-P. Hsieh, "U-air: When urban air quality inference meets big data," in Proc. 19th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining (KDD) New York, NY, 2013, pp. 1436–1444.
[12] H.-P. Hsieh, S.-D. Lin, and Y. Zheng, "Inferring air quality for station location recommendation based on urban big data," in Proc. 21st ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining (KDD), New York, NY, 2015, pp. 437–446.
[13] J. Schmidhuber, "Deep learning in neural networks: An overview," Neural Netw., vol. 61, pp. 85–117, Jan. 2015, doi: 10.1016/j.neunet.2014.09.003.
[14] K. Cho et al., "Learning phrase representations using RNN encoder-decoder for statistical machine translation," arXiv preprint arXiv:1406.1078, 2014.
[15] A. Creswell, T. White, V. Dumoulin, K. Arulkumaran, B. Sengupta, and A. A. Bharath, "Generative adversarial networks: An overview," IEEE Signal Process. Mag., vol. 35, no. 1, pp. 53–65, Jan. 2018, doi: 10.1109/MSP.2017.2765202.
[16] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Proc. Adv. Neural Inf. Process. Syst. (NIPS), New York, NY, 2012, pp. 1097–1105.
[17] L. Gonog and Y. Zhou, "A review: Generative adversarial networks," in 2019 14th IEEE Conf. Ind. Electron. Appl. (ICIEA), Xi'an, China, 2019, pp. 505–510.
[18] Z. Qi, T. Wang, G. Song, W. Hu, X. Li, and Z. Zhang, "Deep air learning: Interpolation, prediction, and feature analysis of fine-grained air quality," IEEE Trans. Knowl. Data Eng., vol. 30, no. 12, pp. 2285–2297, Dec. 2018, doi: 10.1109/TKDE.2018.2823740.
[19] B. Eravci and H. Ferhatosmanoglu, "Diverse relevance feedback for time series with autoencoder based summarizations," IEEE Trans. Knowl. Data Eng., vol. 30, no. 12, pp. 2298–2311, Mar. 2018, doi: 10.1109/TKDE.2018.2820119.
[20] J. Wang and G. Song, "A deep spatial-temporal ensemble model for air quality prediction," Neurocomputing, vol. 314, pp. 198–206, Nov. 2018, doi: 10.1016/j.neucom.2018.06.049.
[21] D. Zhuang, G. Li, and A. Wong, "Discovery of temporal associations in multivariate time series," IEEE Trans. Knowl. Data Eng., vol. 26, pp. 2969–2982, Dec. 2014, doi: 10.1109/TKDE.2014.2310219.
[22] P. W. Soh, J. W. Chang, and J. W. Huang, "Adaptive deep learning-based air quality prediction model using the most relevant spatial-temporal relations," IEEE Access, vol. 6, pp. 38186–38199, Jun. 2018, doi: 10.1109/ACCESS.2018.2849820.
[23] Q. Zhou, H. Jiang, J. Wang, and J. Zhou, "A hybrid model for PM2.5 forecasting based on ensemble empirical mode decomposition and a general regression neural network," Sci. Total Environ., vol. 496, pp. 264–274, Oct. 2014, doi: 10.1016/j.scitotenv.2014.07.051.
[24] L. A. Díaz-Robles et al., "A hybrid ARIMA and artificial neural networks model to forecast particulate matter in urban areas: The case of Temuco, Chile," Atmospheric Environ., vol. 42, no. 35, pp. 8331–8340, Nov. 2008, doi: 10.1016/j.atmosenv.2008.07.020.
[25] X. Zhou et al., "Probabilistic dynamic causal model for temporal data," in 2015 Int. Joint Conf. Neural Netw. (IJCNN), Killarney, Ireland, 2015, pp. 1–8.
[26] S. Deleawe, J. Kusznir, B. Lamb, and D. J. Cook, "Predicting air quality in smart environments," J. Ambient Intell. Smart Environ., vol. 2, no. 2, pp. 145–152, Jan. 2010, doi: 10.3233/ais-2010-0061.
[27] Y. Zheng, L. Capra, O. Wolfson, and H. Yang, "Urban computing: Concepts, methodologies, and applications," ACM Trans. Intell. Syst. Technol., vol. 5, no. 3, pp. 1–55, Sep. 2014, doi: 10.1145/2629592.
[28] X. Yi, J. Zhang, Z. Wang, T. Li, and Y. Zheng, "Deep distributed fusion network for air quality prediction," in Proc. 24th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining (KDD), New York, NY, 2018, pp. 965–973.
[29] Y. Zheng et al., "Forecasting fine-grained air quality based on big data," in Proc. 21st ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining (KDD), New York, NY, 2015, pp. 2267–2276.
[30] M. Längkvist, L. Karlsson, and A. Loutfi, "A review of unsupervised feature learning and deep learning for time-series modeling," Pattern Recognit. Lett., vol. 42, pp. 11–24, Jun. 2014, doi: 10.1016/j.patrec.2014.01.008.
[31] J. Yang, M. N. Nguyen, P. P. San, X. L. Li, and S. Krishnaswamy, "Deep convolutional neural networks on multichannel time series for human activity recognition," in Proc. Int. Joint Conf. Artif. Intell. (IJCAI), New York, NY, 2015, pp. 3995–4001.
[32] X. Li, L. Peng, Y. Hu, J. Shao, and T. Chi, "Deep learning architecture for air quality predictions," Environ. Sci. Pollut. Res., vol. 23, no. 22, pp. 22408–22417, Nov. 2016, doi: 10.1007/s11356-016-7812-9.
[33] B. T. Ong, K. Sugiura, and K. Zettsu, "Dynamically pre-trained deep recurrent neural networks using environmental monitoring data for predicting PM2.5," Neural Comput. Appl., vol. 27, no. 6, pp. 1553–1566, Aug. 2016, doi: 10.1007/s00521-015-1955-3.
[34] J. Zhang, Y. Zheng, and D. Qi, "Deep spatio-temporal residual networks for citywide crowd flows prediction," in Proc. AAAI Conf. Artif. Intell. (AAAI), New York, NY, 2017, pp. 1655–1661.
[35] Z. Wu, X. Wang, Y. G. Jiang, H. Ye, and X. Xue, "Modeling spatial-temporal clues in a hybrid deep learning framework for video classification," in Proc. 23rd ACM Int. Conf. Multimedia (ACM MM), New York, NY, 2015, pp. 461–470.
[36] Y. Sun, X. Wang, and X. Tang, "Hybrid deep learning for face verification," IEEE Trans. Pattern Anal. Mach. Intell., vol. 38, no. 10, pp. 1489–1496, Dec. 2013, doi: 10.1109/TPAMI.2015.2505293.
[37] S. Du, T. Li, Y. Yang, and S. J. Horng, "Deep air quality forecasting using hybrid deep learning framework," IEEE Trans. Knowl. Data Eng., vol. 33, no. 6, pp. 2412–2424, Nov. 2021, doi: 10.1109/TKDE.2019.2954510.
[38] S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural Comput. Appl., vol. 9, no. 8, pp. 1735–1780, Nov. 1997, doi: 10.1162/neco.1997.9.8.1735.
[39] A. Graves, S. Fernández, and J. Schmidhuber, "Bidirectional LSTM networks for improved phoneme classification and recognition," in Artif. Neural Netw. Formal Models Their Appl. – ICANN 2005, Berlin, Heidelberg, 2005, pp. 799–804.
[40] A. Graves and J. Schmidhuber, "Framewise phoneme classification with bidirectional LSTM and other neural network architectures," Neural Netw., vol. 18, no. 5, pp. 602–610, Jul. 2005, doi: 10.1016/j.neunet.2005.06.042.
[41] F. Yu and K. V, "Multi-scale context aggregation by dilated convolutions," in Proc. Int. Conf. Learn. Representations (ICLR), San Juan, Puerto Rico, 2016, pp. 1–13.
[42] Q. Zhang, V. O. Li, J. C. K. Lam, and Y. Han, "Deep-AIR: A hybrid CNN-LSTM framework for fine-grained air pollution forecast," arXiv preprint arXiv:2001.11957, 2020.
[43] X. Zhen, R. Chakraborty, N. Vogt, B. B. Bendlin, and V. Singh, "Dilated convolutional neural networks for sequential manifold-valued data," Proc. IEEE Int. Conf. Comput. Vision, vol. 2019, pp. 10620–10630, Oct.-Nov. 2019, doi: 10.1109/iccv.2019.01072.
[44] Donnelly, A., Misstear, B., and Broderick, B, "Real-time air quality forecasting using integrated parametric and nonparametric regression techniques," Atmospheric Environment, vol. 103, pp. 53-65, 2015.
[45] Schmidhuber J, "Deep learning in neural networks: An overview," Neural networks, vol. 61, pp. 85-117, 2015.
[46] Karpathy A, and Li F F, "Deep visual-semantic alignments for generating image descriptions," in Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3128-3137, 2015.
[47] Zhou Q, Jiang H, Wang J, et al, "A hybrid model for PM2.5 forecasting based on ensemble empirical mode decomposition and a general regression neural network," The Science of the Total Environment, vol. 496, pp.264-274, 2014.
[48] Ong B T, Sugiura K, and Zettsu K, "Dynamically pre-trained deep recurrent neural networks using environmental monitoring data for predicting PM2.5," Neural Computing and Applications, vol. 27, no. 6, pp. 1553-1566, 2016.
[49] Lipton Z C, Berkowitz J, and Elkan C, "A critical review of recurrent neural networks for sequence learning," arXiv preprint arXiv: 1506.00019, 2015.
[50] Gamboa J C B, "Deep learning for time-series analysis," arXiv preprint arXiv:1701.01887, 2017.
[51] S. J. Horng, P. W. Chiang, and W. Zhou, "Time-efficient Dilated Convolution Framework for Region-based Air Quality Prediction," Technical Report, Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, 2020.
[52] P. W. Chiang and S. J. Horng, "Hybrid Time-series Framework for Daily-based PM2.5 Forecasting," IEEE Access, vol. 9, July 2021, pp. 104162–104176, doi: 10.1109/ACCESS.2021.3099111.
