smoker_l1o.m

来自「支持向量机作为统计学习理论的实现方法」· M 代码 · 共 32 行

% SMOKER_L1O Leave One Out validation for SVM trained by SMO.
% [l1o_err, margin, error] = smoker_l1o( K, labels, C, eps, tol, verb )
%
% mandatory inputs:
%  K [N x N ] kernel matrix of N training patterns.
%  labels [1 x N] pattern labels (1 for 1st class, 2 for 2nd class ).
%  C [real] or [2 x real] one trade-off constant for both the classes
%    or two constants for the first and the second class.
%
% optional inputs:
%  eps [real] tolerance of KKT-conditions fulfilment (default 0.001).
%  tol [real] minimal change of optimized Lagrangeians (default 0.001).
%  verb [int] if 1 or 2 then the progress is displayed.
%
%  mandatory outputs:
%   l1o_err [real] leave one out error.
%  
% optional outputs:
%  margin [real] margin of the classifier of the whole problem.
%  error  [real] training error of the whole problem.
%
% See also SMO, SVMCLASS, SVM.

% Statistical Pattern Recognition Toolbox, Vojtech Franc, Vaclav Hlavac
% (c) Czech Technical University Prague, http://cmp.felk.cvut.cz
% Written Vojtech Franc (diploma thesis) 02.11.1999
%
% Modifications:
%  21-Oct-2001, V.Franc
%  18-Oct-2001, V.Franc, created 
%