Study on Growth Model of Tongue Sole Based on Parallel BP Neural Network

Abstract

Fish growth model is an important method to study the growth and development of fish, aiming at the problems of poor performance and low efficiency in processing high-dimensional and large-capacity data in the existing fish growth equations, proposed a growth model of tongue sole built by the parallel BP neural network based on mini-batch gradient descent. This paper introduces the data preprocessing method, details introduction of the parallel BP neural network based on mini-batch gradient descent model design, and uses traditional BP neural network, parallel BP neural network based on stochastic gradient descent and parallel BP neural network based on mini-batch gradient descent build a growth model of tongue sole to compare and analyze. The results show that the RMSE of the tongue sole growth model built by the parallel BP neural network based on mini-batch gradient descent is 1.43, which has a better fitting effect.