Description of slcorrectrate

Home > sltoolbox > perfeval > slcorrectrate.m

slcorrectrate

PURPOSE

SLCORRECTRATE Computes the correct rate of classification

SYNOPSIS

function cr = slcorrectrate(scores, clabels, qlabels, op, varargin)

DESCRIPTION

SLCORRECTRATE Computes the correct rate of classification

 $ Syntax $
   - cr = slcorrectrate(scores, clabels, qlabels, op, ...) 

 $ Arguments $
   - scores:           the scores to support the classification
   - clabels:          the labels of classes
   - qlabels:          the groundtruth of the labels of query samples
   - op:               the option of the score
   - cr:               the correct rate of the score-based classification

 $ Description $
   - cr = slcorrectrate(scores, clabels, slabels, op, ...) 
     Computes the classification correct rate. Suppose we want to 
     classify n samples into m classes, then scores will be an m x n 
     matrix, with the entry at i-th row, j-th column representing the 
     score of the j-th sample in the i-th class. 
     For op, it can take either of the two values: 'high' and 'low'. 
     If op is 'high' then it means that the sample will be classified 
     to the class where it has the highest score value, vice versa 
     for 'low'.
 
 $ History $
   - Created by Dahua Lin on Jun 10th, 2005
   - Modified by Dahua Lin on May 1st, 2005
     - To base on the sltoolbox v4.
   - Modified by Dahua Lin on Aug 9th, 2006
     - Extract slclassify as independent function and base on it
   - Modified by Dahua Lin on Aug 16th, 2006
     - Based on new slclassify to support multiple schemes

CROSS-REFERENCE INFORMATION

This function calls:

• raise_lackinput RAISE_LACKINPUT Raises an error indicating lack of input argument
• slclassify SLCLASSIFY Classifies a set of samples according to final scores

This function is called by:

SOURCE CODE

0001 function cr = slcorrectrate(scores, clabels, qlabels, op, varargin)
0002 %SLCORRECTRATE Computes the correct rate of classification
0003 %
0004 % $ Syntax $
0005 %   - cr = slcorrectrate(scores, clabels, qlabels, op, ...)
0006 %
0007 % $ Arguments $
0008 %   - scores:           the scores to support the classification
0009 %   - clabels:          the labels of classes
0010 %   - qlabels:          the groundtruth of the labels of query samples
0011 %   - op:               the option of the score
0012 %   - cr:               the correct rate of the score-based classification
0013 %
0014 % $ Description $
0015 %   - cr = slcorrectrate(scores, clabels, slabels, op, ...)
0016 %     Computes the classification correct rate. Suppose we want to
0017 %     classify n samples into m classes, then scores will be an m x n
0018 %     matrix, with the entry at i-th row, j-th column representing the
0019 %     score of the j-th sample in the i-th class.
0020 %     For op, it can take either of the two values: 'high' and 'low'.
0021 %     If op is 'high' then it means that the sample will be classified
0022 %     to the class where it has the highest score value, vice versa
0023 %     for 'low'.
0024 %
0025 % $ History $
0026 %   - Created by Dahua Lin on Jun 10th, 2005
0027 %   - Modified by Dahua Lin on May 1st, 2005
0028 %     - To base on the sltoolbox v4.
0029 %   - Modified by Dahua Lin on Aug 9th, 2006
0030 %     - Extract slclassify as independent function and base on it
0031 %   - Modified by Dahua Lin on Aug 16th, 2006
0032 %     - Based on new slclassify to support multiple schemes
0033 %
0034 
0035 %% parse and verify input arguments
0036 
0037 if nargin < 4
0038     raise_lackinput('slcorrectrate', 4);
0039 end
0040 
0041 
0042 %% Make decision
0043 
0044 decisions = slclassify(scores, clabels, op, varargin{:});
0045 
0046 %% Evaluate correct rate
0047 
0048 qlabels = qlabels(:)';
0049 cr = sum(decisions == qlabels) / length(qlabels);
0050 
0051 
0052 
0053 
0054 
0055 
0056

---

Generated on Wed 20-Sep-2006 12:43:11 by m2html © 2003