Comparison of the Efficiency of Classification Methods

• Activity Number: 532
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Rui Wang* and Yoonkyung Lee+
• Companies: The Ohio State University and The Ohio State University
• Keywords: Boosting ; Classification ; Error Rate ; LDA ; SVM ; Efficiency

Abstract

Classification is an important statistical problem with a wide range of applications. A variety of statistical tools have been developed for learning a classification rule from data. Understanding of their relative merits and comparisons help users to choose a proper method in practice. This paper focuses on comparison of model-based classification methods in statistics with algorithmic methods in machine learning in terms of the error rate. Extending Efron's comparison of logistic regression with the LDA under the normal setting, we contrast the support vector machine and boosting with the LDA and logistic regression and study their relative efficiencies based on the limiting behavior of the classification boundary of each method. In addition to the theoretical study, we carry out numerical experiments for more comprehensive comparison under different settings than the normal setting.