ALGORITHM 2: AdaBoost.PL.V2
	Input: Training sets of M participators
	$\{D_{n^{\left(1\right)}}^1,D_{n^{\left(2\right)}}^2,\cdots ,D_{n^{\left(M\right)}}^M\}$, and
	Number of boosting iterations (T).
	Output: The final classifiers H(·)
1	foreach p in M participators parallel do
2	Hp(·) ← AdaBoost$(D_{n^{\left(p\right)}}^p,T)$;
3	Hp*(·) ← the weak classifiers in Hp(·) sorted w.r.t. αp(t);
4	Send Hp*(·) to central agent;
5	end
6	Initialize $\Lambda =\left\{{\lambda }_p=\frac{n^{\left(p\right)}}{{\sum }_{m=1}^M{n}^{\left(m\right)}}\mid p=1,2,\dots ,M\right\}$;
7	for t ← 1 to T do
8	h(t)(·) ← Merge (h1*(t)(·), …, hM*(t)(·));
9	${\alpha }^t\leftarrow {\sum }_{p=1}^M{\alpha }^{p^{\ast }\left(t\right)}{\lambda }_p$;
10	end
11	return $H(\cdot )={\sum }_{t=1}^T{\alpha }^t{h}^{\left(t\right)}(\cdot )$