摘要
研究了人工智能辅助诊断的支持向量机模型,构造了支持向量机疾病确诊模型,设计了症状规范化、从识别域可信度知识规则到SVM分类器训练数值样本的转移方法、样本预处理、SVM模型构造、训练、确诊的过程及方法,以羊为例开展模型和专家的对比实验。实验数据表明,SVM方法能获得85%以上的诊断正确率,具备较好诊断效果。
A SVM diagnosing model is constructed, and a method of standardizing clinic presentation and of the translating CF rules into SVM classifier training numeric vectors is proposed, and the corresponding algorithm and course of diagnosing, sample preprocessing, training to solving is designed. Some contrast diagnosis experiments in the case of local goat disease between the model and human experts are done. Statistics demonstrate that support vector machine method has accurate rate over 85% and a favorable effect.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第5期1796-1799,共4页
Computer Engineering and Design
基金
湖南省自然科学基金项目(10JJ5063)
湖南文理学院自科类重点基金项目(JJZD1002)
关键词
支持向量机
训练
辅助诊断
特征提取
识别域
SVM
training
assistant diagnosing
character extraction
classifying area
