Spam filtering based on AdaBoost and active learning methods

Traditional spam classification has a phenomenon of over-fitting. The AdaBoost method can avoid overfitting situations arise, and by the weighted number of soft classifiers can improve classification performance. To obtain the true labels may require major eort and incur excessive costs (Zliobaite & Bifet 2014). We reduce the number of labeling by active learning method (Cohn 1994, Attenberg 2011).