Table 2.

	Algorithm 3: MAP-DP (generalized algorithm)
Input	x1, …, xN: data ϵ > 0: convergence threshold N0: prior count θ0: prior hyper parameters
Output	z1, …, zN: cluster assignments K: number of clusters
1	K = 1, zi = 1 for all i ∈ 1, …, N
2	Enew = ∞
3	repeat
4	Eold = Enew
5	for i ∈ 1, …, N
6	for k ∈ 1, …, K
7	Update cluster hyper parameters ${\theta }_k^{-i}$ (see (S1 Material))
8	$d_{i,k}=-\text{ln\hspace{0.17em}}f\left(x_i|{\theta }_k^{-i}\right)$
9	di,K+1 = −ln f(xi|θ0)
10	$z_i={arg\; min}_{k\in 1,\dots ,K+1}[d_{i,k}-\text{ln\hspace{0.17em}}{N}_k^{-i}]$
11	if zi = K + 1
12	K = K + 1
13	$E_{\text{new}}={\sum }_{k=1}^K{\sum }_{i:z_i=k}{d}_{i,k}-K\text{\hspace{0.17em}ln\hspace{0.17em}}{N}_0-{\sum }_{k=1}^{K}log\Gamma \left(N_k\right)$
14	until Eold − Enew < ϵ