I'm studying on Pattern Recognition, especially AdaBoost. For this purpose, I read "Experiments with a New Boosting Algorithm" Y. Freund, R. Schapire paper and understood AdaBoost.M1 algorithm. But I have a few questions about AdaBoost.M2.
"Call WeakLearn, providing it with mislabel distribution Dt.". What does this sentence mean? I want to use Euclidean Linear Discriminant classifier with AdaBoost. But, how can I give the distribution to the weak learner? In Adaboost.M1, because of all examples have one value, it didn't cause any problem. But, in Adaboost.M2 an example have K-1 value (corresponding with mislabels).

Please help me.

Murat.