基于偏序结构理论的知识发现方法

摘要

Knowledge discovery in big data is one of key problems for big data application. From the view of highly abstract of cognition, a generalized method of knowledge discovery based on the theory of attribute partial ordered structure diagram (APOSD) and object partial ordered structure diagram (OPOSD) is proposed in this paper. Then, the analysis and compare has been discussed for structural relation of data features in aspects of structure of groups, structure of subgroups, branches and nodes of partial ordered structure diagram. In APOSD, objects with certain characteristics have congregated into several clusters, which express the common attributes of the objects. Using the unique attribute of objects, it can distinguish specific object from the others quickly and effec-tively. Finally, taking the compatibility of traditional Chinese medicine problem as an example, compatibility rules for prescription of apoplexy are analyzed, and application of knowledge discovery in study of compatibility of traditional Chinese medicine is di-scussed in this paper, and the method is proved to be effective and effective and practical by the pre-experiment results.%大数据中的知识发现是大数据应用中的核心热点。本文从高度抽象认知事物视角出发，以表征事物普遍性为特征的概念驱动与表征事物特异性为特征的数据驱动两种方法学为哲学原理，提出了基于属性偏序结构图和对象偏序结构图的知识发现方法。分别从群结构、子群结构、支路、节点等角度对数据特征之间的结构关系进行讨论分析。属性偏序结构图将数据中具有某些共同特征的对象聚类到一起，是数据共性的表达；对象偏序结构图中，通过数据的独有属性可以快速有效的将特异性对象区分于其他对象。最后，以中医药方剂配伍研究问题为例，对张锡纯治疗中风的32个处方进行数据挖掘和知识发现，证明了该方法的有效性和实用性，为大数据知识发现研究提供了新的思路和方法。