Python與MATLAB小練習：計算準確度Accuracy
作者：凱魯嘎吉 - 博客園 http://www.cnblogs.com/kailugaji/
分別使用Python與MATLAB編程，計算聚類準確度。思路爲：首先利用匈牙利算法將訓練後的標籤進行調整，然後再計算準確度。
1. Python程序
1 # Python demo 2 # -*- coding: utf-8 -*- 3 # Author：凱魯嘎吉 Coral Gajic 4 # https://www.cnblogs.com/kailugaji/ 5 # Python小練習：計算準確度Accuracy 6 # 先用匈牙利算法調整標籤，然後再計算準確度 7 import numpy as np 8 # 已經調整過標籤了 9 def cluster_acc(y_true, y_pred): 10 y_true = y_true.astype(np.int64) 11 assert y_pred.size == y_true.size 12 D = max(y_pred.max(), y_true.max()) + 1 13 w = np.zeros((D, D), dtype=np.int64) 14 for i in range(y_pred.size): 15 w[y_pred[i], y_true[i]] += 1 16 from sklearn.utils.linear_assignment_ import linear_assignment 17 # 匈牙利算法調整標籤 18 ind = linear_assignment(w.max() - w) 19 return sum([w[i, j] for i, j in ind]) * 1.0 / y_pred.size 20 21 y_true = np.array([2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1]) 22 y_pred_1 = np.array([1, 1, 2, 1, 1, 2, 2, 2, 3, 2, 2, 3, 1, 3, 3, 2, 3]) # 未調整的標籤 23 y_pred_2 = np.array([2, 2, 3, 2, 2, 3, 3, 3, 1, 3, 3, 1, 2, 1, 1, 3, 1]) # 調整後的標籤 24 result_1 = cluster_acc(y_true, y_pred_1) 25 result_2 = cluster_acc(y_true, y_pred_2) 26 print('1：', result_1) 27 print('2：', result_2)
結果：
1： 0.6470588235294118 2： 0.6470588235294118
2. MATLAB程序
%% MATLAB demo
clear
clc
y_true = [2 2 2 2 2 2 3 3 3 3 3 3 1 1 1 1 1 ];
y_pred_1 = [1 1 2 1 1 2 2 2 3 2 2 3 1 3 3 2 3];
results = Evaluate(y_true,y_pred_1);
fprintf('未調整標籤的準確度：%f\n', results(1));
% --------------------------------------------------
% 實際採用下面這個：先用匈牙利算法對標籤進行調整，然後再計算準確度Accuracy
y_pred_2 = label_map(y_pred_1, y_true);
results = Evaluate(y_true,y_pred_2);
fprintf('調整標籤後的準確度：%f\n', results(1));
%% MATLAB實例：Munkres指派算法 - 凱魯嘎吉 - 博客園
% 來自：https://www.cnblogs.com/kailugaji/p/11765596.html
function [assignment,cost] = munkres(costMat)
% MUNKRES Munkres Assign Algorithm
%
% [ASSIGN,COST] = munkres(COSTMAT) returns the optimal assignment in ASSIGN
% with the minimum COST based on the assignment problem represented by the
% COSTMAT, where the (i,j)th element represents the cost to assign the jth
% job to the ith worker.
%
% This is vectorized implementation of the algorithm. It is the fastest
% among all Matlab implementations of the algorithm.
% Examples
% Example 1: a 5 x 5 example
%{
[assignment,cost] = munkres(magic(5));
[assignedrows,dum]=find(assignment);
disp(assignedrows'); % 3 2 1 5 4
disp(cost); %15
%}
% Example 2: 400 x 400 random data
%{
n=5;
A=rand(n);
tic
[a,b]=munkres(A);
toc
%}
% Reference:
% "Munkres' Assignment Algorithm, Modified for Rectangular Matrices",
% http://csclab.murraystate.edu/bob.pilgrim/445/munkres.html
% version 1.0 by Yi Cao at Cranfield University on 17th June 2008
assignment = false(size(costMat));
cost = 0;
costMat(costMat~=costMat)=Inf;
validMat = costMat<Inf;
validCol = any(validMat);
validRow = any(validMat,2);
nRows = sum(validRow);
nCols = sum(validCol);
n = max(nRows,nCols);
if ~n
return
end
dMat = zeros(n);
dMat(1:nRows,1:nCols) = costMat(validRow,validCol);
%*************************************************
% Munkres' Assignment Algorithm starts here
%*************************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% STEP 1: Subtract the row minimum from each row.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
dMat = bsxfun(@minus, dMat, min(dMat,[],2));
%**************************************************************************
% STEP 2: Find a zero of dMat. If there are no starred zeros in its
% column or row start the zero. Repeat for each zero
%**************************************************************************
zP = ~dMat;
starZ = false(n);
while any(zP(:))
[r,c]=find(zP,1);
starZ(r,c)=true;
zP(r,:)=false;
zP(:,c)=false;
end
while 1
%**************************************************************************
% STEP 3: Cover each column with a starred zero. If all the columns are
% covered then the matching is maximum
%**************************************************************************
primeZ = false(n);
coverColumn = any(starZ);
if ~any(~coverColumn)
break
end
coverRow = false(n,1);
while 1
%**************************************************************************
% STEP 4: Find a noncovered zero and prime it. If there is no starred
% zero in the row containing this primed zero, Go to Step 5.
% Otherwise, cover this row and uncover the column containing
% the starred zero. Continue in this manner until there are no
% uncovered zeros left. Save the smallest uncovered value and
% Go to Step 6.
%**************************************************************************
zP(:) = false;
zP(~coverRow,~coverColumn) = ~dMat(~coverRow,~coverColumn);
Step = 6;
while any(any(zP(~coverRow,~coverColumn)))
[uZr,uZc] = find(zP,1);
primeZ(uZr,uZc) = true;
stz = starZ(uZr,:);
if ~any(stz)
Step = 5;
break;
end
coverRow(uZr) = true;
coverColumn(stz) = false;
zP(uZr,:) = false;
zP(~coverRow,stz) = ~dMat(~coverRow,stz);
end
if Step == 6
% *************************************************************************
% STEP 6: Add the minimum uncovered value to every element of each covered
% row, and subtract it from every element of each uncovered column.
% Return to Step 4 without altering any stars, primes, or covered lines.
%**************************************************************************
M=dMat(~coverRow,~coverColumn);
minval=min(min(M));
if minval==inf
return
end
dMat(coverRow,coverColumn)=dMat(coverRow,coverColumn)+minval;
dMat(~coverRow,~coverColumn)=M-minval;
else
break
end
end
%**************************************************************************
% STEP 5:
% Construct a series of alternating primed and starred zeros as
% follows:
% Let Z0 represent the uncovered primed zero found in Step 4.
% Let Z1 denote the starred zero in the column of Z0 (if any).
% Let Z2 denote the primed zero in the row of Z1 (there will always
% be one). Continue until the series terminates at a primed zero
% that has no starred zero in its column. Unstar each starred
% zero of the series, star each primed zero of the series, erase
% all primes and uncover every line in the matrix. Return to Step 3.
%**************************************************************************
rowZ1 = starZ(:,uZc);
starZ(uZr,uZc)=true;
while any(rowZ1)
starZ(rowZ1,uZc)=false;
uZc = primeZ(rowZ1,:);
uZr = rowZ1;
rowZ1 = starZ(:,uZc);
starZ(uZr,uZc)=true;
end
end
% Cost of assignment
assignment(validRow,validCol) = starZ(1:nRows,1:nCols);
cost = sum(costMat(assignment));
end
%% MATLAB實例：爲匹配真實標籤，對訓練得到的標籤進行調整 - 凱魯嘎吉 - 博客園
% 來自：https://www.cnblogs.com/kailugaji/p/11771226.html
function [ new_label ] = label_map( label, gnd )
%爲匹配真實標籤，對標籤重新調整
K = length(unique(gnd));
cost_mat = zeros(K,K);
for i=1:K
idx = find(label==i);
for j=1:K
cost_mat(i,j) = length(find(gnd(idx)~=j));
end
end
[assignment, ~] = munkres(cost_mat);
[assignedrows, ~]=find(assignment');
new_label = label;
for i=1:K
idx = find(label==i);
new_label(idx) = assignedrows(i);
end
end
%% MATLAB聚類有效性評價指標（外部 成對度量） - 凱魯嘎吉 - 博客園
% 來自：https://www.cnblogs.com/kailugaji/p/11926253.html
function result = Evaluate(real_label,pre_label)
% This fucntion evaluates the performance of a classification model by
% calculating the common performance measures: Accuracy, Sensitivity,
% Specificity, Precision, Recall, F-Measure, G-mean.
% Input: ACTUAL = Column matrix with actual class labels of the training
% examples
% PREDICTED = Column matrix with predicted class labels by the
% classification model
% Output: EVAL = Row matrix with all the performance measures
% https://www.mathworks.com/matlabcentral/fileexchange/37758-performance-measures-for-classification
idx = (real_label()==1);
p = length(real_label(idx));
n = length(real_label(~idx));
N = p+n;
tp = sum(real_label(idx)==pre_label(idx));
tn = sum(real_label(~idx)==pre_label(~idx));
fp = n-tn;
fn = p-tp;
tp_rate = tp/p;
tn_rate = tn/n;
accuracy = (tp+tn)/N; %準確度
sensitivity = tp_rate; %敏感性：真陽性率
specificity = tn_rate; %特異性：真陰性率
precision = tp/(tp+fp); %精度
recall = sensitivity; %召回率
f_measure = 2*((precision*recall)/(precision + recall)); %F-measure
gmean = sqrt(tp_rate*tn_rate);
Jaccard=tp/(tp+fn+fp); %Jaccard係數
result = [accuracy sensitivity specificity precision recall f_measure gmean Jaccard];
end
結果：
未調整標籤的準確度：0.294118 調整標籤後的準確度：0.647059
完成。