[decision tree] use decision tree to diagnose breast cancer

【 Decision tree 】 Use decision tree to diagnose breast cancer

1. Data sets

Wisconsin breast cancer dataset

2. Data set composition ： Healthy tissue ( Benign ) And cancerous tissue ( Malignant ) Digital image of

3. Mission ：

Two classification → Classify tissue samples as benign and malignant .

4. Start

4.1.1 Loading data sets

from sklearn import datasets
data = datasets.load_breast_cancer()

data.data.shape

Data set co ownership 569 Zhang image ,30 Different characteristics .

Mission objectives ：

4.1.2 Data set partitioning

import sklearn.model_selection as ms
X_train, X_test, y_train, y_test = ms.train_test_split(data.data, data.target, test_size=0.2, random_state=42)

Use here 0.2 namely 20% As a test set .

4.2 Build decision tree

Use scikit-learn Of tree modular .

• Don't use any parameters

from sklearn import tree
dtc1 = tree.DecisionTreeClassifier()
dtc1.fit(X_train,y_train)

View the score of the model on the training data set ：

Check the error of the model on the test set ：

See what the decision tree looks like ：

with open("tree.dot", 'w') as f:
    f = tree.export_graphviz(dtc1, out_file=f,
                             feature_names=data.feature_names,
                             class_names=data.target_names)

Even if no parameters are set , Thanks to the sklearn Very good default parameters , Also obtained on the test set **94.7%** The accuracy of .