Deep learning-based research on a long-term time-series prediction algorithm for urban air quality

DingJun Zhang; MingChao Liao; ZhiQiang Huang; LaLao Gao

doi:10.1117/12.2656588

20 October 2022 Deep learning-based research on a long-term time-series prediction algorithm for urban air quality

DingJun Zhang, MingChao Liao, ZhiQiang Huang, LaLao Gao

Proceedings Volume 12451, 5th International Conference on Computer Information Science and Application Technology (CISAT 2022); 1245151 (2022) https://doi.org/10.1117/12.2656588
Event: 5th International Conference on Computer Information Science and Application Technology (CISAT 2022), 2022, Chongqing, China

Abstract

To overcome the limitations of deep learning models that can only capture single temporal feature: First. Concentrating solely on the abstract time series features discovered over the entire dataset, rather than the local periodic segments that contribute more to prediction accuracy; Second, only the time-series variation pattern is fitted, while other non-time series periodic patterns in the PM2.5 data set are omitted, making it impossible to increase the forecast accuracy further. A Bi-directional long and short-term memory-convolutional neural network model including the Attention Mechanism is proposed as BiLSTM-Attention-CNN. The temporal characteristics acquired by BiLSTM filtered by attention mechanism are fused with the cross-periodic features retrieved by 1DCNN to generate an ordered complementary but non-interfering feature. The final experiments revealed that, when compared to other mainstream models, the model reduces error and improves prediction accuracy, and performs well in the long-period time series prediction problem.

Citation Download Citation

DingJun Zhang, MingChao Liao, ZhiQiang Huang, and LaLao Gao "Deep learning-based research on a long-term time-series prediction algorithm for urban air quality", Proc. SPIE 12451, 5th International Conference on Computer Information Science and Application Technology (CISAT 2022), 1245151 (20 October 2022); https://doi.org/10.1117/12.2656588

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available