Upper ocean thermal structure prediction method based on deep belief network

Abstract

The invention discloses an upper ocean thermal structure prediction method based on a deep belief network. The prediction method comprises the following steps of 1) determining proper environment parameters as input factors and predication values of a thermal structure prediction model; 2) establishing an upper ocean thermal structure sample data set, dividing sample data into training data and test data, and performing unified data preprocessing; 3) establishing the deep belief network, and performing non-monitoring pre-training on the sample data layer by layer to obtain relatively excellent parameters of the model preliminarily; 4) based on a back propagation algorithm, performing fine tuning on the parameters of the model according to marks of the training sample to determine optimal parameters; and 5) applying an established upper ocean thermal structure prediction model to test data, and obtaining a predicated temperature value at a specific depth of the upper ocean from the output layer. By adoption of the upper ocean thermal structure prediction method based on the deep belief network provided by the invention, the problems of single fitting and over-fitting of the conventional method are overcome; and a characteristic relation between the ocean surface layer environment parameters and the upper ocean thermal structure is effectively extracted, so that the upper ocean thermal structure prediction accuracy is improved.

Description

technical field [0001] The invention relates to the field of ocean data processing, in particular to a method for predicting upper ocean thermal structure based on a deep belief network. Background technique [0002] The thermal structure of the upper ocean is one of the important indicators to characterize the thermal state of the ocean and the heat exchange process between the ocean and the atmosphere, and is an important part of the ocean environment. Changes in the thermal structure of the upper ocean are closely related to ENSO phenomena, monsoon outbreaks, and typhoon activities. At the same time, changes in it can also cause changes in seawater density, resulting in changes in sea level. Therefore, in-depth study and prediction of the thermal structure changes in the upper ocean can help humans better understand the impact of the ocean on climate, and has important scientific significance for disaster prevention and mitigation. [0003] However, at present, the exist...

AI Technical Summary

Problems solved by technology

[0004] Traditional research methods include dynamical methods and statistical analysis methods. The dynamical methods are affected by ocean observation conditions and temporal and spatial resolutions, and can only roughly characterize the thermal structure of the ocean. It is often difficult to obtain detailed information on the thermal structure of the upper ocean. The analysis method establishes the linear regression relationship between several variables of the sea surface and the internal thermal structure to achieve the prediction of the thermal structure of the upper ocean. This linear statistical mapping method is relatively simple and practical. The complex relationship of nonlinearity and uncertainty, the simple linear relationship cannot achieve good simulation and approximation effect

Images