Research And Implementation Of Milk Yield Prediction Based On Optimized BP Neural Network

In recent years,the stock of dairy cows in China has increased year by year at the end of the year.Milk and dairy products are more and more welcomed by people,and the demand for milk by residents is also increasing.The modern breeding and raising level of dairy cows have greatly improved the production efficiency of dairy farms and the unit yield level of dairy cows,and improved the quality of milk;In this paper,the model is built and the milk yield is predicted and analyzed by integrating the data in the measurement of cow production performance(DHI).The main work of this paper is as follows:After reading and sorting out a large number of documents,summarizing and analyzing the research status at home and abroad and the traditional milk yield prediction model,focusing on the traditional milk yield prediction model and the development and application of modern dairy farming,and then conducting in-depth exploration and reading on the relevant milk yield prediction,we learned that the relevant work is not perfect,and there is still a lot of room for development and excavation,Then,through the data collection and processing of dairy farms,because there are missing values and abnormal values in the data,the average value and box diagram are used to fill in the missing values and detect the abnormal values.After processing the data,the correlation between its eigenvalues and the predicted label variables is analyzed.It is found that there is a weak correlation between the milk yield data.Aiming at this problem,a milk yield prediction model based on rf-pso-bp is proposed.Firstly,the symbolic regression algorithm is used to construct new features to expand the dimension of data and improve the correlation of features;Secondly,aiming at the problem that BP neural network is easy to fall into local optimization,particle swarm optimization is used to optimize the network weight to effectively improve the representation ability of the model;Then,in order to select the effective features of the data,the random forest is used to sort the importance of the features,and the more important features are selected as the model input to effectively improve the modeling ability.The experiments show that the model proposed in this paper can effectively predict the milk yield;Finally,a cow intelligent management system based on B/S structure is designed and implemented.