【摘要】 以吉林省瘦肉型优质猪为研究样本,讨论了胴体等级评定方法和评定体系,设计了胴体等级预报模型。胴体等级预报采用了多元线性回归和人工神经网络2种方法。多元线性回归方程的准确率为80%,神经网络方法的准确率为90%。猪胴体级别模糊描述考虑了肉质方面的因素,使猪胴体等级的最终评定体系更科学、更合理。该方法适合二元、三元杂交等"杂交猪"胴体等级的评定。更多还原

【Abstract】 <Abstrcat> 　In the paper, appraising methods of pork carcass grade were investigated. Appraisement models for pork carcass grade were devised. The appraisement was done by multiplex linear regression and artificial neural network. The accuracy of multiplex linear regression is 80% and that of the artificial neural network is 90%. The study showed that artificial neural network method was more effective and accurate. The factors of meat quality were calculated in fuzzy appraisement for pork grade, which made the appraisement method more scientific and reasonable. The grading method for carcass of crossbred Jilin swine was suitable to grade the dyadic or triadic crossbred swine.更多还原