ML之RF：基于Matlab利用RF算法实现根据乳腺肿瘤特征向量高精度(better)预测肿瘤的是恶性还是良性

ML之RF：基于Matlab利用RF算法实现根据乳腺肿瘤特征向量高精度(better)预测肿瘤的是恶性还是良性

 

 

目录

输出结果

实现代码

---

 

 

 

 

输出结果

更新……

 

 

实现代码

%RF：RF实现根据乳腺肿瘤特征向量高精度(better)预测肿瘤的是恶性还是良性
 
load data.mat
 
a = randperm(569);
Train = data(a(1:500),:);
Test = data(a(501:end),:);
 
P_train = Train(:,3:end);
T_train = Train(:,2);
 
P_test = Test(:,3:end);
T_test = Test(:,2);
 
model = classRF_train(P_train,T_train); 
 
[T_sim,votes] = classRF_predict(P_test,model);
 
count_B = length(find(T_train == 1));
count_M = length(find(T_train == 2));
total_B = length(find(data(:,2) == 1));
total_M = length(find(data(:,2) == 2));
number_B = length(find(T_test == 1));
number_M = length(find(T_test == 2));
number_B_sim = length(find(T_sim == 1 & T_test == 1));
number_M_sim = length(find(T_sim == 2 & T_test == 2));
disp(['病例总数：' num2str(569)...
      '  良性：' num2str(total_B)...
      '  恶性：' num2str(total_M)]);
disp(['训练集病例总数：' num2str(500)...
      '  良性：' num2str(count_B)...
      '  恶性：' num2str(count_M)]);
disp(['测试集病例总数：' num2str(69)...
      '  良性：' num2str(number_B)...
      '  恶性：' num2str(number_M)]);
disp(['良性乳腺肿瘤确诊：' num2str(number_B_sim)...
      '  误诊：' num2str(number_B - number_B_sim)...
      '  确诊率p1=' num2str(number_B_sim/number_B*100) '%']);
disp(['恶性乳腺肿瘤确诊：' num2str(number_M_sim)...
      '  误诊：' num2str(number_M - number_M_sim)...
      '  确诊率p2=' num2str(number_M_sim/number_M*100) '%']);
   
figure
 
index = find(T_sim ~= T_test);
plot(votes(index,1),votes(index,2),'r*')
hold on
 
index = find(T_sim == T_test);
plot(votes(index,1),votes(index,2),'bo')
hold on
 
legend('红色*是错误分类样本','蓝色空心圆是正确分类样本')
 
plot(0:500,500:-1:0,'r-.')
hold on
 
plot(0:500,0:500,'r-.')
hold on
 
line([100 400 400 100 100],[100 100 400 400 100])
 
xlabel('输出为类别1的决策树棵数')
ylabel('输出为类别2的决策树棵数')
title('随机森林分类器性能分析—Jason niu')   
 
 
Accuracy = zeros(1,20);
for i = 50:50:1000
    i
    accuracy = zeros(1,100);
    for k = 1:100
        model = classRF_train(P_train,T_train,i);
        T_sim = classRF_predict(P_test,model);
        accuracy(k) = length(find(T_sim == T_test)) / length(T_test);
    end
     Accuracy(i/50) = mean(accuracy);
end
 
 
figure
plot(50:50:1000,Accuracy)
xlabel('随机森林中决策树棵数')
ylabel('分类正确率')
title('随机森林中决策树棵数对性能的影响—Jason niu')

 

相关文章
RF：RF实现根据乳腺肿瘤特征向量高精度(better)预测肿瘤的是恶性还是良性