论文部分内容阅读
洪水灾害灾情识别的实质就是建立各个洪水灾害灾情决策指标与洪灾等级之间的非线性关系。鉴于洪灾成因机制的复杂性和发生过程的随机性,在灰色聚类的基础上引入信息熵概念,提出了具有典型指数白化权函数,并采用加权平均原则的灰色信息熵聚类。该方法有效解决了“零权重”问题,通过引用熵权所反应的数据本身的效用值来修正指标的权重系数,充分利用了样本遗留信息,并极大地保留了聚类系数的蕴涵信息。实例证明,灰色信息熵聚类的评价过程直观简单,结果合理有效,能有效扩大灰色聚类在实际工程中的应用范围。
The essence of flood disaster identification is to establish the nonlinear relationship between flood disaster decision indicators and flood level. In view of the complexity of flood mechanism and the randomness of its occurrence process, this paper introduces the concept of information entropy based on gray clustering and proposes gray information entropy clustering with typical exponential whitening weight function and weighted average principle. The method effectively solves the problem of “zero weight ”, amends the weight coefficient of the index by referring to the utility value of the data corresponding to the entropy weight, makes full use of the information left over from the sample and greatly preserves the implication information of the clustering coefficient . The example shows that gray information entropy clustering evaluation process is intuitive and simple, the results are reasonable and effective, which can effectively expand the scope of application of gray clustering in practical engineering.