Research on noise reduction combined model for short-term precipitation forecast

Xiaofei Pang

doi:10.1117/12.2652365

20 October 2022 Research on noise reduction combined model for short-term precipitation forecast

Xiaofei Pang

Proceedings Volume 12350, 6th International Workshop on Advanced Algorithms and Control Engineering (IWAACE 2022); 123500A (2022) https://doi.org/10.1117/12.2652365
Event: 6th International Workshop on Advanced Algorithms and Control Engineering (IWAACE 2022), 2022, Qingdao, China

Abstract

In order to improve the forecast accuracy of daily rainfall, it is convenient for flood control departments to make decisions. Under the condition of abundant meteorological data, a combined BiLSTM precipitation forecast model based on denoising autoencoder is proposed. The combined model is mainly used for training and prediction through noise reduction of input data, feature extraction and distinguishing the importance of meteorological information. The model uses 19 meteorological factors related to daily precipitation (including 20 to 20 hours' cumulative precipitation) as input vector, and the next 24 hours' precipitation as output vector. The results show that the model has the best prediction performance with root mean square error of 13.04 to about 15.18mm in the study area except for the three stations closest to the sea. The three cities closest to the sea in the study area have achieved the best prediction results by using the DBNPF model, and the root mean square error is 17.83 to about 18.95mm. The experimental results show that the combined model proposed in this paper is feasible and provides a new idea for daily rainfall prediction.

Citation Download Citation

Xiaofei Pang "Research on noise reduction combined model for short-term precipitation forecast", Proc. SPIE 12350, 6th International Workshop on Advanced Algorithms and Control Engineering (IWAACE 2022), 123500A (20 October 2022); https://doi.org/10.1117/12.2652365

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