摘要
信息的不完备、不确定是复杂决策环境中不可避免的问题。从不完备、不确定的海量信息中发现某种特定目标的潜在有用的知识,若没有一种降低不确定性、变复杂为简的有效处理海量数据集的方法,而直接利用相应的海量数据进行挖掘,不仅数据处理效率低,更由于传统的方法基本是以不可分辨关系为基础的,因此,其变复杂为简的方法要求形成互不相交的等价数据子集,这在实际处理中不易满足。文中给出商空间的保序性的有关概念,给出求商(拟)半序直接方法及其数学证明,然后以序关系代替传统的不可分辨关系,形成逼近某种特定目标的子集序列,从而获取潜在有用的知识。
There are some questions that cannot be avoided in complex decision making environment, such as imperfect and uncertainty information. If there is no method that can reduce the nondeterminacy and make complicated thing to simple in dealing with mass data set usefully, but using the matching mass data directly to find out the potential information with some special target from the imperfect and uncertainty mass data. If you do in this way, you will find it is inefficient. And the traditional method is based on the tmdistinguishability, so the method that makes the complicated thing to simple requires to form a mutually disjointed equivalent data subset, but it is hard to meet in reality. Firstly, this paper offers the relevant concept about quotient space, then provides the direct method to obtain the quotient semi - order and its mathematical proof, at last in order to get the potential useful information, it uses the order relation to replace the traditional one that is undistinguished and forms a sequence of subsets which is close to some special target.
出处
《计算机技术与发展》
2007年第9期71-74,共4页
Computer Technology and Development
基金
国家自然科学基金资助项目(60675031)
安徽省自然科学基金资助项目(05042000208)
安徽省教育厅重点自然科学研究项目(2006KJ015A)
关键词
商空间
保序性
等价关系
半序
quotient space
holding order property
equivalence relation
half order
