Algorithm 1: TNNCPD

Input: $\mathcal{T}$ Output: $\tilde{\mathcal{T}},A_i,Y_i,i=1,2,3$ 1: Initialize: $\tilde{\mathcal{T}}=\mathcal{T},$ $A_i=$ R order vectors in front of the of $T_{(i)}$; The largest number of iterations: $n=1000$; convergence threshold: $\eta =0.001$; 2: while no convergence do 3:  for $i=1$ to 3 do 4:    $Y_i={\mathrm{\Theta }}_{\frac{{\gamma }_i}{{\alpha }_i+{\beta }_i}}(\frac{{\alpha }_i{T}_{(i)}+{\beta }_i{\tilde{T}}_{(i)}}{{\alpha }_i+{\beta }_i})$ 5:  end for $i=1$ to 3 do 6:   ${\tilde{T}}_i={(\frac{{\displaystyle \sum _{j=1}^3}{\beta }_j{y}_j}{{\displaystyle \sum _{j=1}^3}{\beta }_j})}_{\left(i\right)}$ 7:  end for   $A_1={\tilde{T}}_{(1)}{\left[{(A_2\odot A_3)}^T\right]}^{†},A_2={\tilde{T}}_{(2)}{\left[{(A_3\odot A_1)}^T\right]}^{†},A_3={\tilde{T}}_{(3)}{\left[{(A_2\odot A_1)}^T\right]}^{†}$ 8: end