摘要
针对中药方剂功效归纳问题,提出了一种基于人工神经网络新的高维数据归约方法。新方法利用属性间先验的相似信息,得到属性距离矩阵,然后将矩阵引入神经网络,通过训练神经网络得到最终数据归约结果。依据这个方法实现了一个中药方剂分析系统。实验表明,新方法在中药方剂功效的自动归纳中获得很好的效果。
A novel reduction method of high dimensions based on artificial neural network was proposed for the effect reduction of Chinese traditional medicine prescription. Based on the s/milarity informat/on between properties, it created a property distance matrix. The matrix was loaded into an artificial neural network, and through training the artificial neural network, it got the reduce result of high dimensions. Based on the new method, a prescription analyze system of Chinese traditional medicine was implemented. The efflcieney of the new method was proved by experiments in effect reduction of Chinese traditional medicine.
出处
《四川大学学报（工程科学版）》
EI
CAS
CSCD
北大核心
2006年第1期92-97,共6页
Journal of Sichuan University (Engineering Science Edition)
基金
国家自然科学基金(60473071)
四川省重点科技计划项目(04SG1640)
高等学校博士学科点专项科研基金SRFDP(20020610007)
国家中医药管理局基金SATCM资助项目(2003JP40)
关键词
神经网络
高维数据归约
相似度
方剂功效
矢量
相似距离
ANN
high dimensions reduction
similarity
prescription effect
vector
similarity distance
