Input: Training set $D=\left\{\left(x^1,y^1\right),\cdots ,\left(x^n,y^n\right)\right\}$

Output: Regression chain model $h_j=X\to R$

1. generate $D_1$

2. $D_1^{*}=\left\{\left(x^1,y_1^1\right),\cdots ,\left(x^n,y_1^n\right)\right\}$

3. For $j=1$ to m

4. $h_j=D_j^{*}\to R$

5. If $j<m$

6. Generating data sets $D_{j+1}^{*}$

7. $h_j^i=D_j^{*}\backslash D_j^{*i}\to R$

8. for $x_j^{*i}\in D_j^{*i}$

9. $x_{j+1}^{*i}=x_j^{*i}$

10. $j^i=h_j^i\left(x_j^{*i}\right)$

11. $x_{j+1}^{*i}=\left[x_j^{*i},{\widehat{y}}_j^i\right]$

12. $D_{j+1}^{*}=D_{j+1}^{*}\cup \left(x_{j+1}^i,y_{j+1}^i\right)$