神经网络结构优化在肉鸡生长性能预测中的应用  

Application of the Neural Network structural optimization in broiler growth performance prediction

作　　者：吴广发[1] 林丕源[1] 黄沛杰[1] 肖媚燕[1] 张键锋[1] 

机构地区：[1]华南农业大学信息学院,广东广州510642

出　　处：《广东农业科学》2011年第22期126-129,共4页Guangdong Agricultural Sciences

基　　金：广东省科技计划项目(2007A020300010;2010B020315024);广东高校优秀青年创新人才培养计划项目(LYM09034);广东省大学生科技创新项目(1056410038)

摘　　要：动物生长性能预测是养殖业信息技术的核心应用之一。分析季节性因素对肉鸡生长状况的影响,进而建模和预测肉鸡生长性能,可以为大规模肉鸡养殖提供良好的决策支持。基于神经网络模型及其结构优化理论,采用国内著名肉鸡养殖公司提供的案例数据,进行系统的神经网络结构优选试验,得到了较为稳定和有效的预测模型。Animal growth performance prediction was one of core applications of the information technology in breeding industry. By analyzing the seasonal influence on broiler breeding and making further modeling and prediction, modern poultry breeding companies could get better decision support for large scale poultry raising. Based on the Neural Network and its structural optimization theory and the broiler growth dataset provided by the most famous poultry raising company in China, this paper deployed systematic experiments on the structural optimization of Neural Network, and achieved a relatively stable and effective prediction model.

关 键 词：性能预测 神经网络 结构优化 肉鸡养殖 季节性因素 

分 类 号：TB183[一般工业技术]