We consider the problem of classifying documents not by topic, but by overall sentiment, e.g., determining whether a review is positive or negative. Using movie reviews as data,...
Introduction to support vector learning roadmap. Part 1 Theory: three remarks on the support vector method of function estimation, Vladimir Vapnik generalization performance of ...
Feature selection is an important problem for pattern classification systems. We study how to select good features according to the maximal statistical dependency criterion base...
From the Publisher: How does it differ from first edition? Includes new material on: * support vector machines (SVM's), * fat shattering dimensions * applications to neural net...
We compare discriminative and generative learning as typied by logistic regression and naive Bayes. We show, contrary to a widelyheld belief that discriminative classiers are al...
Social media, news, blog, policy document mentions