CART : Classification And Regression Trees

Matlab'ta CART : Classification And Regression Trees islemlerinin nasil kullanildigi hakkinda basit bir kac uygulamayi makale iceriginde bulabilirsiniz. Ayni zamanda CART : Classification And Regression Trees islemlerinde datasetlerin nasil cagirilabilecegi hakkinda da bilgi sahibi olabileceksiniz. Uygulama 1 : Breast Cancer Wisconsin (Original) Data Set kullanilarak classification ve regration tree olusturuluyor. Matlab Kod clear all; close all; clc; dataset = load('breast-cancer-wisconsin.data'); train = dataset(:,1:10); class = dataset(:,11); classificationTree = fitctree(train,class) view(classificationTree) view(classificationTree,'mode','graph') regressionTree = fitrtree(train,class); view(regressionTree) view(regressionTree,'mode','graph') Uygulama 2 : Breast Cancer Wisconsin (Original) Data Set kullanilarak dogruluk, hata oranlari ve confusion matrix degerleri hesaplaniyor. Matlab Kod clear all; close all; clc; dataset = load('breast-cancer-wisconsin.data'); dataEgitim = dataset(1:600,1:10); dataTest = dataset(601:683,1:10); classEgitim = dataset(1:600,11); classTest = dataset(601:683,11); tree = ClassificationTree.fit(dataEgitim, classEgitim) t = classregtree(dataEgitim, classEgitim); cvv = crossval(tree); error = kfoldLoss(cvv) dogruluk = 1 - error c1 = tree.predict(dataTest); cMat = confusionmat(classTest, c1) error =    0.0517 dogruluk =    0.9483 cMat =    67     2 0    14 Uygulama 3 :Matlab Kod clear all; close all; clc; x1 = <0 1 0 1 0 1 0 1>'; x2 = <0 0 0 0 1 1 1 1>'; x3 = <0 0 1 1 0 0 1 1>'; inData = ; outData = <'-', '-', '+', '+', '+', '+', '-', '-'>'; mytree = treefit(inData, outData, 'method', 'classification', 'splitmin', 2, 'prune', 'on', 'splitcriterion', 'gdi') treedisp(mytree); Decision tree for classification 1  if x1<0.5 then node 2 elseif x1>=0.5 then node 3 else - 2  if x2<0.5 then node 4 elseif x2>=0.5 then node 5 else - 3  if x2<0.5 then node 6 elseif x2>=0.5 then node 7 else - 4  if x3<0.5 then node 8 elseif x3>=0.5 then node 9 else - 5  if x3<0.5 then node 10 elseif x3>=0.5 then node 11 else - 6  if x3<0.5 then node 12 elseif x3>=0.5 then node 13 else - 7  if x3<0.5 then node 14 elseif x3>=0.5 then node 15 else - 8  class = - 9  class = + 10  class = + 11  class = - 12  class = - 13  class = + 14  class = + 15  class = - Keyifli Calismalar Dilerim.