Procedure 1 Build PM
1:	$\mathrm{T}\leftarrow \left\{\right\}$	▷ Start with an empty train set
2:	for c in C do	▷ For each class
3:	$dat{a}_t\leftarrow subset({\tau }_t,c)$	▷ ${\tau }_t$ is the target user’s train set
4:	$dat{a}_{other}\leftarrow subset(instances\left(U_{other}\right),c)$
5:	$dat{a}_{all}\leftarrow dat{a}_t\cup dat{a}_{other}$
6:	Cluster $dat{a}_{all}$ using k-means for $k=2,\dots ,UpperBound$ and select the optimal k according to some clustering quality index.
7:	$S\leftarrow$${arg\; max}_{g\in G}\left|dat{a}_t\cap g\right|$	▷ G is the set of the resulting k groups
8:	$\mathrm{T}\leftarrow \mathrm{T}\cup S\cup dat{a}_t$
9:	end for
10:	$weight\left(\mathrm{T}\right)$    ▷ Assign a weight to each instance such that the importance of ${\tau }_t$ increases as more training data of the target user is available.
11:	Build model using training set $\mathrm{T}$.