福州大学智能技术课件-6-bagging-boosting.pdfVIP

Part I: Assessment of Accuracy of Classification Rules/Classifiers Part II: Ensemble methods Bagging Boosting D.Y.YE智能技术课程 1 IF-THEN Rules for Classification • Represent the knowledge in the form of IF-THEN rules R: IF hair = gold AND eyes = blue THEN grouping = 1 – Rule antecedent/precondition vs. rule consequent • Assessment of a rule: coverage and accuracy – ncovers = # of tuples (elememts, records, instances) covered by R – ncorrect = # of