Using Adaboost with SVM for classification

Question

I know that Adaboost tries to generate a strong classifier using a linear combination of a set of weak classifiers.

However, I've read some papers suggesting Adaboost and SVMs work in harmony (even though SVM is a strong classifier) in certain conditions and cases.

I'm not able to grasp from an architectural and programming perspective how they work in combination. I've read many papers (maybe the wrong ones) which didn't explain clearly how they work together.

Can someone throw some light on how they work in a combination for effective classification? Pointers to some papers / articles / journals would also be appreciated.

Answer 1

This paper is quite good. It simply says that SVM can be treated as a weak classifier if you use fewer samples to train it (let's say less than half of the training set). The higher the weights the more chance it will be trained by the 'weak-SVM'

edit: link fixed now.

Answer 2

The paper AdaBoost with SVM-based component classifiers by Xuchun Li etal also gives an intuition.
In a short but maybe biased summary: they are trying to make svm classifiers "weak"(slightly over 50%) by tuning the parameters to avoid the cases one classifier may have too much weight or all the classifiers fire similarly.