Scikit Learn - K近邻分类器



此分类器名称中的K代表k个最近邻,其中k是由用户指定的整数。因此,顾名思义,此分类器实现基于k个最近邻的学习。k值的选取取决于数据。让我们借助一个实现示例来进一步理解它。

实现示例

在这个例子中,我们将使用scikit-learn的KNeighborsClassifier在名为Iris Flower数据集的数据集上实现KNN。

  • 此数据集每个不同的鸢尾花物种(setosa、versicolor、virginica)有50个样本,即总共150个样本。

  • 对于每个样本,我们有4个特征,分别命名为萼片长度、萼片宽度、花瓣长度、花瓣宽度。

首先,导入数据集并打印特征名称,如下所示:

from sklearn.datasets import load_iris
iris = load_iris()
print(iris.feature_names)

输出

['sepal length (cm)', 'sepal width (cm)', 'petal length (cm)', 'petal width (cm)']

示例

现在我们可以打印目标,即表示不同物种的整数。这里0 = setosa,1 = versicolor,2 = virginica。

print(iris.target)

输出

[
   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
]

示例

以下代码行将显示目标的名称:

print(iris.target_names)

输出

['setosa' 'versicolor' 'virginica']

示例

我们可以使用以下代码行检查观察值和特征的数量(iris数据集有150个观察值和4个特征):

print(iris.data.shape)

输出

(150, 4)

现在,我们需要将数据分成训练数据和测试数据。我们将使用Sklearn的train_test_split函数将数据按70(训练数据)和30(测试数据)的比例分割:

X = iris.data[:, :4]
y = iris.target
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.30)

接下来,我们将使用Sklearn预处理模块进行数据缩放,如下所示:

from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)

示例

以下代码行将给出train和test对象的形状:

print(X_train.shape)
print(X_test.shape)

输出

(105, 4)
(45, 4)

示例

以下代码行将给出新的y对象的形状:

print(y_train.shape)
print(y_test.shape)

输出

(105,)
(45,)

接下来,从Sklearn导入KNeighborsClassifier类,如下所示:

from sklearn.neighbors import KNeighborsClassifier

为了检查准确性,我们需要导入Metrics模型,如下所示:

from sklearn import metrics
We are going to run it for k = 1 to 15 and will be recording testing accuracy, plotting it, showing confusion matrix and classification report:
Range_k = range(1,15)
scores = {}
scores_list = []
for k in range_k:
   classifier = KNeighborsClassifier(n_neighbors=k)
   classifier.fit(X_train, y_train)
   y_pred = classifier.predict(X_test)
   scores[k] = metrics.accuracy_score(y_test,y_pred)
   scores_list.append(metrics.accuracy_score(y_test,y_pred))
result = metrics.confusion_matrix(y_test, y_pred)
print("Confusion Matrix:")
print(result)
result1 = metrics.classification_report(y_test, y_pred)
print("Classification Report:",)
print (result1)

示例

现在,我们将绘制K值与相应的测试精度之间的关系。这将使用matplotlib库完成。

%matplotlib inline
import matplotlib.pyplot as plt
plt.plot(k_range,scores_list)
plt.xlabel("Value of K")
plt.ylabel("Accuracy")

输出

Confusion Matrix:
[
   [15 0 0]
   [ 0 15 0]
   [ 0 1 14]
]
Classification Report:
            precision   recall   f1-score    support
         0     1.00     1.00        1.00        15
         1     0.94     1.00        0.97        15
         2     1.00     0.93        0.97        15

micro avg      0.98     0.98        0.98        45
macro avg      0.98     0.98        0.98        45
weighted avg   0.98     0.98        0.98        45

Text(0, 0.5, 'Accuracy')
Kneighbors Classifier

示例

对于上述模型,我们可以选择K的最优值(由于精度在此范围内最高,因此任何介于6到14之间的值),例如8,并重新训练模型,如下所示:

classifier = KNeighborsClassifier(n_neighbors = 8)
classifier.fit(X_train, y_train)

输出

KNeighborsClassifier(
   algorithm = 'auto', leaf_size = 30, metric = 'minkowski',
   metric_params = None, n_jobs = None, n_neighbors = 8, p = 2,
   weights = 'uniform'
)
classes = {0:'setosa',1:'versicolor',2:'virginicia'}
x_new = [[1,1,1,1],[4,3,1.3,0.2]]
y_predict = rnc.predict(x_new)
print(classes[y_predict[0]])
print(classes[y_predict[1]])

输出

virginicia
virginicia

完整的可运行程序

from sklearn.datasets import load_iris
iris = load_iris()
print(iris.target_names)
print(iris.data.shape)
X = iris.data[:, :4]
y = iris.target

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30)

from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)

print(X_train.shape)
print(X_test.shape)

from sklearn.neighbors import KNeighborsClassifier

from sklearn import metrics

Range_k = range(1,15)
scores = {}
scores_list = []
for k in range_k:
   classifier = KNeighborsClassifier(n_neighbors=k)
   classifier.fit(X_train, y_train)
   y_pred = classifier.predict(X_test)
   scores[k] = metrics.accuracy_score(y_test,y_pred)
   scores_list.append(metrics.accuracy_score(y_test,y_pred))
result = metrics.confusion_matrix(y_test, y_pred)
print("Confusion Matrix:")
print(result)
result1 = metrics.classification_report(y_test, y_pred)
print("Classification Report:",)
print (result1)
%matplotlib inline
import matplotlib.pyplot as plt
plt.plot(k_range,scores_list)
plt.xlabel("Value of K")
plt.ylabel("Accuracy")

classifier = KNeighborsClassifier(n_neighbors=8)
classifier.fit(X_train, y_train)

classes = {0:'setosa',1:'versicolor',2:'virginicia'}
x_new = [[1,1,1,1],[4,3,1.3,0.2]]
y_predict = rnc.predict(x_new)
print(classes[y_predict[0]])
print(classes[y_predict[1]])
广告