Algorithm 1.

AdaBoost.M1

1: Initialize the boosting weights Dn,1 = 1/N for xn ∈ S.

2: for t = 1, …, T do

3:     Train the tth weak classifier ft so as to minimize Jt.

4:     Get estimate of xn ∈ S: ht,n = ft(xn).

5:     Calculate the error of ht,n, εt: ${\epsilon }_t={\displaystyle {\sum }_{n=1}^N}{D}_{t,n}I(h_{t,n}\ne y_n)$

6:     Set βt = εt/(1 − εt).

7:     Update the boosting weights Dt+1,n using Eq.(2).

8: end for

9: return The final classifier H(x).