Prediction method of probability integral parameter based on miv-gp algorithm to optimize bp neural network

Abstract

The invention proposes a combined algorithm (GP) based on genetic algorithm and particle swarm algorithm to optimize the probability integral method parameter prediction model of BP neural network, and adopts mean influence value algorithm (MIV) to optimize the input layer of BP neural network, thereby reducing the The complexity of the network can improve the prediction accuracy. The MIV-GP-BP model was established with the measured data of 50 working faces as the training set and test set of BP neural network, and the accuracy and reliability of the model prediction results were analyzed. All are between 0.0058 and 1.1575, the maximum relative error of q, tanβ, b, θ is less than 5.42%, the average relative error is less than 2.81%, the relative error of s / H is less than 9.66%, and the average relative error is less than 4.31 % (the parameter itself is smaller), the optimized neural network model has higher prediction accuracy and stability.

Description

technical field [0001] The invention relates to the field of probability integral parameter prediction, in particular to a probability integral parameter prediction method for optimizing a BP neural network based on an MIV-GP algorithm. Background technique [0002] With the rapid development of China's economy, the demand for coal resources is also increasing. In recent years, a large number of underground coal resources have been mined, and underground coal mining has caused a series of environmental problems, such as: surface subsidence, cracks, dust, solid waste, etc., which have brought serious threats to the production and life of mining areas. In order to maximize the extraction of coal resources and reduce the large-scale subsidence of the surface, scholars have carried out extensive research on the theory of mining subsidence prediction. Among them, the probability integration method based on the stochastic medium mechanics theory is a widely used mining subsidence...

AI Technical Summary

Problems solved by technology

These algorithms include simulated annealing (SA), GA algorithm, PSO algorithm, etc. Among these optimization algorithms, PSO algorithm is simple and has high operation efficiency, but it has the disadvantages of low precision and easy to fall into local optimum.

Images