Grain Yield Prediction Of RBF Neural Network Based On Improved Particle Swarm Optimization Algorithm

Grain production represents the strength of a country's economic strength,which is very important to the development of a country.Grain production and human production and life are closely related,providing adequate material support to the survival and development of human beings.Grain is the necessities of people's survival,and has a important influence on a country's future and destiny.At present,there are still many countries having the problem of food,and food is still one of the most serious problems facing humanity,which has aroused wide attention from all over the world.As we all know,China is a large agricultural country,but it has a relatively scarce land resources,so to ensure the stability of food security has a profound impact on the stability of China's development.Therefore,it is a very important measure to predict the development of agricultural grain production.This paper mainly studies the use of particle swarm optimization algorithm to improve the RBF neural network,so that the prediction rate is faster and the prediction accuracy is higher,and it is not easy to show local optimal.It continuously updates of the particle position and velocity by adding a local search operator in the particle swarm algorithm,and the fitness parameters is optimal in the global search range.Using particle swarm optimization algorithm gets RBF neural network parameters,which can make the parameters of RBF neural network optimal.And the grain yield prediction is close to the actual values.First,the principle and process of particle swarm algorithm are described,then we can get random initialization of the particle swarm by analysis of particle swarm optimization algorithm.We can get the optimal solution by iterations of the sample output under the condition of initializing the particle swarm.Then the optimization of particle swarm optimization algorithm was used to process parameters of the RBF neural network,and the yield of grain samples is normalized,and the structure of RBF neural network is initialized and the parameters,,i i iw c ? are given random initial value which can get the optimal particle group model to achieve the purpose of this paper to satisfy.Secondly,the RBF neural network model is introduced.The principle of RBF neural network model is introduced,and the principle is analyzed.The paper introduces the origin of the RBF neural network,as well as the characteristic,and briefly introduces the common application of RBF neural network.RBF neural network used in life because of its unique advantages.Through the analysis of RBF neural network model,the prediction model of the Grain Yield is established,then the parameters of the RBF neural network model is optimized.The parameters is optimized by using particle swarm optimization algorithm and the effect of the grain yield prediction is good.Based on a sample of grain yield,the grain production is measured in the next few years,and the prediction result is compared by the actual output,and the error analysis is made.Conclusions can be drawn that the result of grain yield prediction model by using particle swarm optimization algorithm of neural network is close to the actual values.Finally,we optimized the RBF neural network by using particle swarm optimization algorithm and trains grain production samples,and get the optimal parameters of RBF neural network.So we can get the value of grain yield prediction.