Landslide displacement prediction method based on SVR-LSTM mixed deep learning

Abstract

The invention discloses a landslide displacement prediction method based on SVR-LSTM mixed deep learning. The landslide displacement prediction method comprises the steps of 1, performing EMD decomposition on displacement data to obtain an IMF component and a residual term; 2, substituting trend term data into a trend term prediction model SVR for training; and taking residual data as test items and substituting the test items into the trained model to obtain a trend item prediction result; 3, obtaining candidate training attributes of a plurality of periodic terms; 4, calculating mutual information and a Pearson's correlation coefficient and selecting influence factors; 5, using the influence factors as training factors of the LSTM to obtain a plurality of LSTM models; and substituting the residual data into the LSTM model to obtain predicted output values, and adding the predicted output values of the IMF components to obtain a periodic term prediction result; 6, correspondingly adding data in the periodic term prediction result and the trend term prediction result to obtain a total displacement prediction result. The SVR and LSTM methods are adopted to predict the sum, so that the stability and accuracy of the prediction result are greatly improved, the credibility of the prediction result is ensured, and the calculation is efficient.

Description

technical field [0001] The invention belongs to the technical field of landslide displacement prediction, and in particular relates to a landslide displacement prediction method based on SVR-LSTM hybrid deep learning. Background technique [0002] Landslides include landslides, loess landslides, etc., which are common geological disasters. They refer to rocks and loess on slopes that are affected by their own gravity, rainfall, groundwater, earthquakes, surface water immersion, and river water at the bottom of the mountain. Continuous scouring and unreasonable human engineering activities, such as excavating slope toe, blasting mountains, storing (discharging) water in reservoirs, earthquakes, freezing and thawing, etc., under a series of actions, along the through shear failure surface, the overall It is a kind of geological disaster that slides downward along the slope angle in a scattered or scattered manner. The accidents caused by such disasters continue to occur freque...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a landslide displacement prediction method based on SVR-LSTM hybrid deep learning, which is used to solve the time-consuming and labor-intensive establishment of complex mechanical equations and statistical models in the prior art, with a large amount of calculation and a long modeling period , and the problem that traditional modeling uses low data and the accuracy of prediction is not high

Images