基于模糊C-均值聚类的斜坡稳定性判别

The Discretion of Stability of Slope Based on Fuzzy C-means Clustering

采用模糊C-均值聚类方法对斜坡的稳定性进行判别.针对传统的模糊C-均值聚类方法(FCM)在处理此问题时表现出来的极大不稳定性,以样本特征均值代替FCM中随机初始中心,避免了传统FCM对初始中心敏感的缺陷,纠正了其聚类结果对未知斜坡稳定性的判别.改进后的聚类结果能够保持斜坡样本的基本特征属性,可较为准确的判别出斜坡是稳定的,基本稳定的,还是稳定性差的.

Fuzzy c-means clustering was used to discrete the stability of slopes.Aiming at the great instability what usually emerged while dealing with this kind questions with the traditional fuzzy c-means clustering method(FCM),the initial clustering center produced by chance were replaced by the characteristic means of the sample.It avodied the blemish that FCM was sensitive to the initial center.And also,it rectifird the discretion of stability of unknown slopes with the clustering result.The clustering result improved could keep basic characteristic property of the slope sample,and distinguish accurately a slope whether was stable,basic stable or instable.