Wavelet analysis coupled LSTM neural network water quality prediction method

• Main Authors: LU CHENG, LIU SHUJUN, ZHOU GUOLONG, MA ZHANJUN, WANG HENGJIAN, LAN DESHUN, LIANG ZEHUA
• Format: Patent
• Language: Chinese; English
• Published: 08.03.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIRCHEMICAL OR PHYSICAL PROPERTIES; MEASURING; METEOROLOGY; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS; TESTING

A water quality prediction method based on wavelet analysis coupled with an LSTM neural network comprises the following steps: acquiring monitoring data of water quality parameters of a water body to be predicted, and performing wavelet decomposition on the monitoring data to obtain approximate components and detail components of the monitoring data; and inputting the approximate component and the detail component of the monitoring data as input variables into a pre-trained LSTM neural network to obtain a prediction result of the water quality parameters. Because the monitoring data of the water quality parameters are subjected to wavelet decomposition, the time sequence is decomposed into a series of low-frequency and high-frequency components, and new features are generated, and the features can express the time-frequency information of the monitoring data of the water quality parameters, so that the non-stationary part of the data can be better processed, and the information is extracted from the data rich