AN EXTENSION OF A MINIMAX APPROACH TO MULTIPLE CLASSIFICATION

スポンサーリンク

概要

• 論文の詳細を見る

• When mean vectors and covariance matrices of two classes are available in a binary classification problem, Lanckriet et al. [6] propose a minimax approach for finding a linear classifier which minimizes the worst-case (maximum) misclassification probability. In this paper, we extend the minimax approach to a multiple classification problem, where the number m of classes could be more than two. Assume that mean vectors and covariance matrices of all the classes are available, but no further assumptions are made with respect to class-conditional distributions. Then we define a problem for finding linear classifiers which minimize the worst-case misclassification probability α^^-. Unfortunately, no efficient algorithms for solving the problem are known. So we introduce the maximum pairwise misclassification probability β^^- instead of α^^-. It is shown that β^^- is a lower bound of α^^- and a good approximation of α^^- when m or α^^- are small. We define a problem for finding linear classifiers which minimize the probability β^^- and show some basic properties of the problem. Then the problem is transformed to a parametric Second Order Cone Programming problem (SOCP). We propose an algorithm for solving the problem by using nice properties of it. We conduct preliminary numerical experiments and confirm that classifiers computed by our method work very well to benchmark problems.

• 社団法人日本オペレーションズ・リサーチ学会の論文

著者

• Mizuno Shinji

  Tokyo Institute Of Technology

• Nakata Kazuhide

  Tokyo Institute Of Technology

• Kitahara Tomonari

  Tokyo Institute of Technology

• Kitahara Tomonari

  Tokyo Inst. Technol. Tokyo Jpn

• Mizuno Shinji

  Tokyo Inst. Technol. Tokyo Jpn

関連論文

• QUADRATIC AND CONVEX MINIMAX CLASSIFICATION PROBLEMS
• SDPA PROJECT : SOLVING LARGE-SCALE SEMIDEFINITE PROGRAMS(the 50th Anniversary of the Operations Research Society of Japan)
• 2-C-8 Performance Assessment of Philippine Administrative Divisions by Means of Data Envelopment Analysis
• AN EXTENSION OF A MINIMAX APPROACH TO MULTIPLE CLASSIFICATION
• LOWER BOUNDS FOR THE MAXIMUM NUMBER OF SOLUTIONS GENERATED BY THE SIMPLEX METHOD(SCOPE (Seminar on Computation and OPtimization for new Extensions))

スポンサーリンク