Algorithm 1 Constructing a PCC-Tree

Require: A root node $\mathbf{X}={\left\{{\mathbf{x}}_i\right\}}_{i=1}^N$, where $x_i$ is the i th instance with n condition attributes ${\left\{A_k\right\}}_{k=1}^n$ and one decision attribute D; the stopping criterion $\epsilon$. Ensure: A PCC-Tree.  if the samples in X belong to some class then   Mark X as a leaf node and assign the class as its label.   return.  end if  for each attribute $A_k,k=1,2,\cdots ,n$ in X do   for each value $c_i$ in $A_k$ do    Compute the Pearson’s correlation coefficient P of two vectors: $P_{c_j}\left(A_k\right)=P(V(A_k,c_j)$ and $V\left(D\right))$,    where P denotes Pearson’s correlation coefficient and V denotes one vector.   end for   $c_{jk}^{*}=\arg {max}_{c_j}{P}_{c_j}\left(A_k\right)$.  end for  Get the best attribute $A_{k^{*}}$ and the splitting point $c_{k^{*}}$, where $k^{*}=\arg {max}_k{P}_{c_j}\left(A_k\right)$.  Suppose $p\left(X\right)$ is the proportion of samples covered by X.  if $p\left(\mathbf{X}\right)<\epsilon$ then   Mark X as leaf node.   Assign the maximum class of samples in X to this leaf node.   return  else   Split X into two subsets $X_1$ and $X_2$, based on $A_{k^{*}}$ and $c_{k^{*}}$.   if $p\left(X_1\right)==0$ or $p\left(X_2\right)==0$ then    Mark X as a leaf node.    Assign the maximum class of samples in X to this leaf node.    return   end if   Recursively search the new tree nodes from $X_1$ and $X_2$ by Algorithm 1, respectively.  end if