Seasonal Broiler Growth Performance Prediction Based on Observational Study

Abstract

For the modern poultry breeding companies, it is worthwhile to predict the poultry growth performance parameters to help future production and management. Most of the existing literature of broiler growth performance is based on well designed experiment. However, for the farmers who carry on the large scale breeding all the year round, it is impractical to adjust all the ambient factors to suitable levels systematically like in the laboratory. In this case, it requires an observational study based on the massive historical data gradually cumulated in production. In this paper, observational study is proposed for seasonal broiler growth performance prediction. Systematic observational study with statistical analysis and data mining technology is adopted including macro analysis, exploratory data analysis, and modeling. Case study using the broiler growth dataset of the most famous poultry raising company in China shows the effectiveness of our approach.