Algorithm 1 Centralized Pattern Relation Table Algorithm

Input: Training data set D, test sequence $\sigma$ Output: Detect if $\sigma$ is normal or abnormal 1:Extract the event set E and the necessary event set $E_n$. 2:Learn the loop pattern set $LP$ by Subroutines 1 and 2. 3:Split $d_i\in D$ by Subroutine 3. 4:Build the Pattern Relation Table T by Subroutine 4. 5:Detect if $\sigma$ is normal or not by Subroutine 5.