Broiler Growth Performance Analysis: From Correlation Analysis, Multiple Linear Regression, to Neural Network

Abstract

The purpose of this study is to investigate the data fitting for broiler growth performance parameters. In this paper, the gradual advancing analysis methods, from correlation analysis, multiple linear regression, to neural network, are proposed. The mean technology roadmap is: firstly, correlation analysis is used to detect the degree of correlation between the broiler growth performance parameter and the candidate input variables. And then choose the predictor variables that have good correlation with the dependent variable to build the multiple linear regression or neural network prediction model, or both, according to the linear degree of correlations. Combined prediction may be chose once both models have good prediction performances. We use the broiler growth dataset of the most famous poultry raising company in China to evaluate our approach and the results show the effectiveness of our approach.