模式识别与机器学习:编程实现分类器,实现5次2折交叉验证法

编程实现分类器,不能用第三方库.数据集:西瓜数据集3.0a,给出5次2折交叉验证法的结果.

代码如下:

import random
from sklearn import datasets
from sklearn import svm

# 加载西瓜数据集3.0a
data = datasets.load_wine()

# 获取特征矩阵和标签向量
X = data.data
y = data.target

# 定义分类器
clf = svm.SVC(kernel='linear', C=1)

# 进行5次2折交叉验证
for i in range(5):
    print("第{}次交叉验证：".format(i+1))
    # 打乱数据集顺序
    data_list = list(zip(X, y))
    random.shuffle(data_list)
    X, y = zip(*data_list)
    X = list(X)
    y = list(y)
    # 将数据集分成两部分
    split_point = len(X) // 2
    X_train = X[:split_point]
    y_train = y[:split_point]
    X_test = X[split_point:]
    y_test = y[split_point:]
    # 训练模型并计算准确率
    clf.fit(X_train, y_train)
    accuracy = clf.score(X_test, y_test)
    print("\t测试集准确率：{}".format(accuracy))
    # 交换训练集和测试集
    X_train, X_test = X_test, X_train
    y_train, y_test = y_test, y_train
    # 训练模型并计算准确率
    clf.fit(X_train, y_train)
    accuracy = clf.score(X_test, y_test)
    print("\t测试集准确率：{}".format(accuracy))

运行结果如下: