运用主成分分析与聚类分析确定猪繁殖性状的选育指标

Research on Selection of Representative Traits for Reproductive Performance in Pigs by Principal Component Analysis and Cluster Analysis

测定了长白猪的11个繁殖性状,并计算各性状的表型参数和性状间的相关系数,探讨应用主成分分析和聚类分析方法确定长白母猪繁殖性状选育指标。结果表明:11个繁殖性状被综合为4个主成份,且这4个主成分可保留88.312%的信息量,基本反映了长白母猪繁殖性状指标所含的信息。2种分析结果提示:在长白母猪繁殖性状的选育和性能测定的选育指标中,至少应包括断奶窝重、初生个体重、育成率和乳头数4个性状指标。研究结果为长白猪繁殖性状的选育提供了参考依据。

Eleven reproductive traits of Landrace sow were studied.The phenotypic parameter of these traits and the coefficients of correlation among them were computed.The author employed the hierarchical principal component analysis and cluster analysis to select the representative traits or reproductive performance in landrace sows.The results indicated that eleven reproductive traits could be simplified 4 principal components.These 4 principal components could account for 88.312% variance.According to the same results from the two mathematics methods,least four traits including the weaning weight of litter,the survival rate at weaning,the newborn body weight,the number of nipples 4 properties should be considered.The results of principal component analysis and cluster analysis will provide the reference for selection of reproductive traits for Landrace sows.