利用SVM对乳腺癌进行预测

关于SVM的原理可参考https://zhuanlan.zhihu.com/p/24638007

其中的KKT条件和强对偶性的互相推导可参考我之前的博客：https://blog.csdn.net/qq_35985044/article/details/85324714

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as snsdata = pd.read_csv(&＃39;breast_cancer_01/data.csv&＃39;)
print(data.columns)
print(data.head(5))
print(data.describe())

部分结果如下：

Index([&＃39;id&＃39;, &＃39;diagnosis&＃39;, &＃39;radius_mean&＃39;, &＃39;texture_mean&＃39;, &＃39;perimeter_mean&＃39;,&＃39;area_mean&＃39;, &＃39;smoothness_mean&＃39;, &＃39;compactness_mean&＃39;, &＃39;concavity_mean&＃39;,&＃39;concave points_mean&＃39;, &＃39;symmetry_mean&＃39;, &＃39;fractal_dimension_mean&＃39;,&＃39;radius_se&＃39;, &＃39;texture_se&＃39;, &＃39;perimeter_se&＃39;, &＃39;area_se&＃39;, &＃39;smoothness_se&＃39;,&＃39;compactness_se&＃39;, &＃39;concavity_se&＃39;, &＃39;concave points_se&＃39;, &＃39;symmetry_se&＃39;,&＃39;fractal_dimension_se&＃39;, &＃39;radius_worst&＃39;, &＃39;texture_worst&＃39;,&＃39;perimeter_worst&＃39;, &＃39;area_worst&＃39;, &＃39;smoothness_worst&＃39;,&＃39;compactness_worst&＃39;, &＃39;concavity_worst&＃39;, &＃39;concave points_worst&＃39;,&＃39;symmetry_worst&＃39;, &＃39;fractal_dimension_worst&＃39;],dtype=&＃39;object&＃39;)id diagnosis radius_mean texture_mean perimeter_mean area_mean \
0 842302 M 17.99 10.38 122.80 1001.0
1 842517 M 20.57 17.77 132.90 1326.0
2 84300903 M 19.69 21.25 130.00 1203.0
3 84348301 M 11.42 20.38 77.58 386.1
4 84358402 M 20.29 14.34 135.10 1297.0

#columns取列名，index取行名
features_mean = list(data.columns[2:12])
features_se = list(data.columns[12:22])
features_worst = list(data.columns[22:32])
features_worst

[&＃39;radius_worst&＃39;,&＃39;texture_worst&＃39;,&＃39;perimeter_worst&＃39;,&＃39;area_worst&＃39;,&＃39;smoothness_worst&＃39;,&＃39;compactness_worst&＃39;,&＃39;concavity_worst&＃39;,&＃39;concave points_worst&＃39;,&＃39;symmetry_worst&＃39;,&＃39;fractal_dimension_worst&＃39;]

#删除ID列
data.drop("id",axis=1,inplace=True)
#将B良性替换成0， M恶性替换成1
data[&＃39;diagnosis&＃39;] = data[&＃39;diagnosis&＃39;].map({&＃39;B&＃39;:0, &＃39;M&＃39;:1})
corr = data[features_mean].corr()
plt.figure(figsize=(10,10))sns.heatmap(corr, annot=True)
plt.show()

from sklearn.model_selection import train_test_split
from sklearn import svm
from sklearn import metrics
from sklearn.preprocessing import StandardScaler# 特征选择
features_remain = [&＃39;radius_mean&＃39;,&＃39;texture_mean&＃39;, &＃39;smoothness_mean&＃39;,&＃39;compactness_mean&＃39;,&＃39;symmetry_mean&＃39;, &＃39;fractal_dimension_mean&＃39;] # 抽取30%的数据作为测试集，其余作为训练集
train, test = train_test_split(data, test_size = 0.3)# in this our main data is splitted into train and test
# 抽取特征选择的数值作为训练和测试数据
train_X = train[features_remain]
train_y=train[&＃39;diagnosis&＃39;]
test_X= test[features_remain]
test_y =test[&＃39;diagnosis&＃39;]# 采用Z-Score规范化数据，保证每个特征维度的数据均值为0，方差为1
ss = StandardScaler()
#提取训练集数据的均值和方差，并利用这两个参数对训练集进行标准化
train_X = ss.fit_transform(train_X)
#利用训练集的均值和方差对测试集进行标准化
test_X = ss.transform(test_X)model = svm.SVC()
model.fit(train_X,train_y)
prediction = model.predict(test_X)
print(&＃39;准确率: &＃39;, metrics.accuracy_score(prediction,test_y))

准确率: 0.9415204678362573

还有用PCA降维的，可参考https://blog.csdn.net/Vincent_Chu/article/details/90046985