Inputs:	The target dataset XT (the target domain), the number of clusters C, the known cluster centroids v̂i, i = 1, …, C, or the historical dataset XS (the source domain), the specific values of involved parameters in TI-KT-CM or TII-KT-CM, e.g. η, β, and γ, the maximum iteration number maxiter, the termination condition of iterations ε.
Outputs:	The memberships U, the cluster centroids V, and the labels of all patterns in XT.
Extracting knowledge from the source domain:
Setp1:	Generate the historical cluster centroids v̂i(i = 1, …, C) in the source domain XS via other soft-partition clustering methods, e.g., FCM or MEC (Skip this step if the historical cluster centroids v̂i(i = 1, …, C) are given).
Step2:	Compute the historical cluster centroid-based memberships ũij(i = 1, …, C; j = 1, …, N) of all data instances in XT to those historical cluster centroids v̂i(i = 1, …, C) via Eq. (3) or (6).
Performing clustering in the target domain:
Step 1:	Set the iteration counter t=0 and randomly initialize the memberships U(t) which satisfies 0 ≤ uij(t)≤ 1 and ${\sum }_{i=1}^C{u}_{ij}(t)=1$.
Step 2:	For TI-KT-CM, generate the cluster centroids V(t) via Eq.(18), U(t), and v̂i(i = 1, …, C).
	For TII-KT-CM, generate the cluster centroids V(t) via Eq. (20), U(t), v̂i(i = 1, …, C), and ũij(i = 1, …, C; j = 1, …, N).
Step 3:	For TI-KT-CM, calculate the memberships U(t + 1) via Eq. (19), V(t), and v̂i(i = 1, …, C).
	For TII-KT-CM, calculate the memberships U(t+1) via Eq. (21), V(t), and v̂i(i = 1, …, C).
Step 4:	If ∥ U(t + 1) − U(t)∥ < ε or t=maxiter go to Step 5, otherwise, =t+1 and go to Step 2;
Step 5:	Output the eventual cluster centroids V and memberships U in XT, and determine the label of each individual in XT according to U.