Algorithm 2:Test

Input: Data $(x_i,y_i,z_i),i=1,\dots ,n$.             Cluster number M.             Times to repeat K. Output: p-value. 1Find M clusters. 2Estimate the T. 3 for $k\leftarrow 1$ to K do 4     Generate samples with Algorithm 1 5     Estimate $T_i$ on the generated samples. 6 end 7 Compute the p-value: $$p=\frac{1}{K}{\sum }_{t=1}^T\{T_k\ge T\}.$$