BP神经网络在凉山半细毛羊成年生产性状早期预测中的应用研究

用matlab软件建立BP神经网络,通过凉山半细毛羊初生到断奶的5个生产性状(羔羊初生重、羔羊断奶重、初生至断奶日增重、断奶毛长度、断奶毛细度)预测成年(2.5年)的2个生产性状(成年体重、成年剪毛量),并通过预测结果和实测结果的统计分析验证了所构建的BP神经网络的有效性,预测结果和实测结果的回归分析相关系数(R)分别为0.74和0.711,为羔羊的选育和养殖场经济的预算提供依据。

The Application of BP Neural Network to Predict the Traits of 2.5-year Liangshan Semifine-wool Sheep

This study set up a neural network model to predict the two 2.5-year traits of Liangshan semifine-wool sheep(weight and wool clip at 30 months).The five traits(birth weight,weaning weight,daily gain weight from birth to weaning,weaning staple length and weaning staple fibre diameter) were used as inputs to the network.The outputs of the models were the weight and wool clip at 30 months.simulated values obtained with neural network correspond closely to the experimental results.The R of linear regression analysis of simulated values and experimental results was 0.74 and 0.711,respectively.The proposed neural network model can be used in the prediction of Liangshan semifine-wool sheep.