BP人工神经网络在家蚕杂交组合预测中的应用初探

Preliminary Research on the Prediction for Hybrid Combination Performance Based on BP Artificial Neural Network

利用前馈神经网络(BPANN)建立了家蚕杂种一代的6个经济性状的预测模型,通过验证得出,以双亲成绩作为输入神经元,利用BPANN模型预测杂种一代6个经济性状的平均预测误差均在5%左右。表明利用神经网络模型用来预测杂种一代成绩具有研究应用前景。

The prediction models of the six economic traits on silkworm hybrid were established using BP artificial neural network （BP ANN）.It is found by test that use parents as an input neurons the average prediction error of the six economic traits of the hybrid is about 5%. There is prospect of research and application in the silkworm hybrid combination performance.