Algorithm 2 Outlier Detection

Input   •$X_{train}$         % Data to train the RBFN;   •$X_{test}$         % Data to test the RBFN;   •$n\_outlier$         % n° of outliers in training data;   •$\theta$         % Constant of outlier reduction;      Output   •$C_{Outlier}$         % Set of outliers. 1: procedure Outlier_detection($X_{train}$,$X_{test}$,$n\_outlier$,$\theta$) 2: 3:     for $i\in 1:size\left(X_{train}\right)$ do 4:         $po{t}_{trei}\left(i\right)=P_{if}\left(X_{train}\left(i\right)\right)$         % The IP. 5:     end for 6: 7:     $po{t}_{trei}=sort\left(po{t}_{trei}{,}^{\prime }ascen{d}^{\prime }\right)$         % Sort in ascend order. 8:     $\delta =po{t}_{trei}(\theta ·n\_outlier)$         % The threshold. 9: 10:     for $i\in 1:size\left(X_{test}\right)$ do 11:         if $P_{if}\left(X_{test}\left(i\right)\right)\le \delta$ then 12:            $X_{test}\left(i\right)\in C_{Outlier}$ 13:         end if 14:     end for 15: return $C_{Outlier}$ 16: 17: end procedure