03 测试我们的算法
本例使用datasets数据集中的鸢尾花数据集
import numpy as np import matplotlib.pyplot as plt from sklearn import datasets # 加在鸢尾花数据集 iris = datasets.load_iris() X = iris.data y = iris.target X.shape # Out[4]: # (150, 4) y.shape # Out[5]: # (150,)
train_test_split
平时我们在拿到一个数据集时，往往将其一部分用于对机器进行训练，另一部分用于对训练过后的机器进行测试，即train_test_split
y """ Out[6]: array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]) """ shuffle_indexes = np.random.permutation(len(X)) # 打乱数据，形成150个索引的随机排列 shuffle_indexes """ Out[8]: array([ 22, 142, 86, 111, 72, 80, 17, 137, 5, 66, 33, 55, 40, 122, 108, 24, 45, 110, 68, 46, 118, 44, 136, 121, 78, 31, 103, 35, 105, 107, 76, 116, 84, 144, 123, 57, 42, 7, 38, 28, 117, 115, 89, 58, 126, 74, 49, 27, 94, 77, 85, 21, 119, 132, 100, 120, 6, 104, 62, 53, 64, 41, 106, 26, 29, 18, 129, 146, 148, 1, 82, 139, 135, 96, 127, 56, 37, 130, 65, 149, 113, 92, 131, 2, 4, 125, 54, 79, 50, 61, 112, 95, 19, 109, 102, 141, 30, 39, 83, 25, 140, 60, 12, 20, 138, 71, 59, 11, 13, 0, 52, 91, 3, 73, 23, 124, 15, 14, 81, 97, 75, 114, 16, 69, 32, 134, 36, 8, 63, 51, 147, 67, 93, 47, 133, 48, 143, 43, 34, 98, 87, 88, 145, 70, 90, 9, 10, 128, 101, 99]) """ test_ratio = 0.2 # 设置测试数据集的占比 test_size = int(len(X) * test_ratio) test_indexes = shuffle_indexes[:test_size] # 测试数据集索引 train_indexes = shuffle_indexes[test_size:] # 训练数据集索引 X_train = X[train_indexes] y_train = y[train_indexes] X_test = X[test_indexes] y_test = y[test_indexes] print(X_train.shape) print(y_train.shape) # (120, 4) (120,) print(X_test.shape) print(y_test.shape) # (30, 4) (30,)
sklearn中的train_test_split
sklearn中同样为我们提供了将数据集分成训练集与测试集的方法
# 首先创建一个kNN分类器my_knn_clf，略 # 导入模块 from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=666) print(X_train.shape) print(y_train.shape) # (120, 4) (120,) print(X_test.shape) print(y_test.shape) # (30, 4) (30,) my_knn_clf.fit(X_train, y_train) # Out[32]: # KNN(k=3) y_predict = my_knn_clf.predict(X_test) sum(y_predict == y_test)/len(y_test) # Out[34]: # 1.0
04 分类准确度
本例使用datasets中手写识别数据集来演示分类准确度的计算
import numpy as np import matplotlib import matplotlib.pyplot as plt from sklearn import datasets digits = datasets.load_digits() # 手写识别数据集 digits.keys() # Out[3]: # dict_keys(['data', 'target', 'target_names', 'images', 'DESCR']) print(digits.DESCR) X = digits.data # 数据集 y = digits.target # 标记 # 随便取一个数据集 some_digit = X[666] some_digit_image = some_digit.reshape(8, 8) plt.imshow(some_digit_image, cmap = matplotlib.cm.binary) plt.show()
执行结果：
scikit-learn中的accuracy_score
# 导入split方法先将数据集拆分 from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2) # 导入创建kNN分类器的方法 from sklearn.neighbors import KNeighborsClassifier knn_clf = KNeighborsClassifier(n_neighbors=3) knn_clf.fit(X_train, y_train) y_predict = knn_clf.predict(X_test) # 导入sklearn中计算准确度的方法 from sklearn.metrics import accuracy_score accuracy_score(y_test, y_predict) # Out[27]: # 0.99444444444444446
