In order to study the development degree of spheric weathered granite in Zhuhai

146 survey projects were selected

and the classification indicators were statistically analyzed.The classification of the development degree of spheric weathered granite was conducted according to the Guangdong Provincial Standard.Meanwhile

the system clustering method and K-means clustering method are applied to analyze the revealing rate

line development rate

and multiple revealing rate.The results show that the optimal number of clusters in clustering methods is different from the standard trichotomy method

and the threshold values cannot be directly compared.To establish a correspondence between clustering and classification

the three original indicators were compressed into two principal components through principal component analysis

obtaining clusters with optimal classifications of 3 and 6.Finally

based on the clustering results of the principal components

we proposed suggested values for the classification thresholds.