matlab-(含教程)基于BP神经网络的乳腺癌识别算法的MATLAB仿真,包含数据库,输出识别率,召回率以及F1

clc; clear all; close all; warning off; addpath 'func\'; rng('default'); load DATA\breast.mat %训练比例 L = 0.6; P = breast(:,1:9); T = round(breast(:,end)/2); MTKL= 100; I1 = find(T==1); I2 = find(T==2); idx1= [I1(1:floor(L*length(I1)));I2(1:floor(L*length(I2)))]; idx2= [I1(1+floor(L*length(I1)):end);I2(1+floor(L*length(I2)):end)]; for j = 1:MTKL j rng(j); Ptrain = P(idx1,:); Ptest = P(idx2,:); Ttrain = T(idx1); Ttest = T(idx2); %数据输入 net2 = newff(Ptrain',Ttrain',4,{'tansig', 'purelin'}, 'traingd'); % 隐含层有5个神经元 net2.trainParam.goal = 1e-5; net2.trainParam.epochs = 1000; net2.trainParam.lr = 0.0025; net2.trainParam.showWindow = 0; net2 = train(net2,Ptrain',Ttrain'); y = [round(sim(net2,Ptest'))]'; [A,~]= confusionmat(Ttest,y); %计算-1类的评价值 c1_precise(j) = A(1,1)/(A(1,1) + A(2,1)); c1_recall(j) = A(1,1)/(A(1,1) + A(1,2)); c1_F1(j) = 2 * c1_precise(j) * c1_recall(j)/(c1_precise(j) + c1_recall(j)); %计算1类的评价值 c2_precise(j) = A(2,2)/(A(1,2) + A(2,2)); c2_recall(j) = A(2,2)/(A(2,1) + A(2,2)); c2_F1(j) = 2 * c2_precise(j) * c2_recall(j)/(c2_precise(j) + c2_recall(j)); end idx=[]; for i =1:MTKL if isnan(c1_precise(i))==1 | isnan(c1_recall(i))==1 | isnan(c1_F1(i))==1 | isnan(c2_precise(i))==1 | isnan(c2_recall(i))==1 | isnan(c2_F1(i))==1 idx=[idx,i]; end end c1_precise(idx)=[]; c1_recall(idx)=[]; c1_F1(idx)=[]; c2_precise(idx)=[]; c2_recall(idx)=[]; c2_F1(idx)=[]; R = [mean(c1_precise),mean(c1_recall),mean(c1_F1),mean(c2_precise),mean(c2_recall),mean(c2_F1)] figure; bar([mean(c1_precise),mean(c2_precise)]);title('识别率'); figure; bar([mean(c1_recall),mean(c2_recall)]);title('召回率'); figure; bar([mean(c1_F1),mean(c2_F1)]);title('F1'); save R1.mat c1_precise c1_recall c1_F1 c2_precise c2_recall c2_F1