论文部分内容阅读
基于模糊C均值(FCM)聚类算法建立了实测洪水过程的模糊聚类模型,模型可将实测样本分为若干类,每一类的聚类中心即为一个典型洪水过程.由于传统FCM聚类算法易陷入局部极值点,难以适应洪水过程分类具有的数据量大、维数较高的特点,因此采用遗传算法对其进行了改进.应用改进算法对一个实例进行聚类分析,并结合基于可能性定理的聚类有效性准则,对聚类结果作进一步的有效性评价.分析表明:改进算法产生的分类结果比较合理,较接近于实际情况,可以应用于洪水过程分类.
Based on fuzzy C-means clustering (FCM) clustering algorithm, a fuzzy clustering model is established for the measured flood process. The model can divide the measured samples into several types, and the clustering center of each type is a typical flood process. Because traditional FCM clustering The algorithm is easy to fall into the local extreme point and it is difficult to adapt to the large amount of data and high dimensionality of the flood process classification.Then the genetic algorithm is used to improve the algorithm.The clustering analysis of an example is made with the improved algorithm, The validity criterion of clustering validity theorem is used to evaluate the validity of the clustering results. The analysis shows that the classification results obtained by the improved algorithm are more reasonable and closer to the actual situation and can be applied to flood process classification.