论文部分内容阅读
何湘藩在《模糊分级聚类分析方法》一文中给出了对多对象多指标进行按预期类别分级聚类的系统评价方法,在此基础上,又提出了“优度分析法”,其基本思想是:在分级聚类的基础上,应用综合优度来对被归入同一类别的对象进行排序.但是,用这样的方法解决问题时,还会有两个(或多个)对象并列的机会.因而,该文试图通过:①对原方法中的定义指标优度量;②确定上层指标下对象的隶属度两个步骤的改造来实现更精确的聚类排列.同时,也使“优度分析方法”在多层指标下更方便于应用.最后给出了改造后的方法的计算机流程框图。
He Xiangfan gave a systematic evaluation method of multi-object and multi-index clustering according to the expected classification in the article of “fuzzy hierarchical clustering analysis method”. On this basis, he also proposed the “excellentness analysis method” and its basic idea Is: based on hierarchical clustering, the use of comprehensive goodness of the object to be classified in the same category are sorted.But, in this way to solve the problem, there will be two (or more) object juxtaposed opportunities .Therefore, the paper attempts to: (1) to measure the optimal metric of the original method, (2) to determine the degree of membership of the upper target, to achieve a more accurate clustering transformation, and Method "is more convenient to apply under the multi-level index.Finally, the computer flow chart of the modified method is given.