【摘要】 [目的]研究1种基于SVM二叉树的猪呼吸系统疾病诊断方法。[方法]首先利用聚类思想,定义了1种基于疾病特征权重的分离程度;每次将分离程度最大的类分离出来,得到了累积误差更小的决策树;并将二叉树SVM用于常见的4种呼吸系统疾病的诊断试验。[结果]该方法可行,可用于猪呼吸系统疾病的早期诊断。[结论]为促进我国猪养殖业的健康发展及农民增收提供了参考。更多还原

【Abstract】 [Objective]To study a diagnostic approach to diseases in pig’s respiratory system based upon SVM binary tree. [Method] First with the help of clustering theory,the degree of separation based upon the characteristics of diseases is defined. Each time the type of highest degree of separation is isolated to get a decision tree with smaller accumulated errors and SVM binary tree is applied in the diagnostic experiment of four common respiratory diseases. [Result] The method is practicable and can be applied in the diagnosis of pig’s respiratory diseases at the early stage. [Conclusion] It provides references to the healthy development of pig husbandry in China and increases in breeders’ incomes.更多还原