利用鳶尾花數據集繪制P-R曲線

#利用鳶尾花數據集繪制P-R曲線
print(__doc__)      #打印注釋
 
import matplotlib.pyplot as plt
import numpy as np
from sklearn import svm, datasets
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import average_precision_score
from sklearn.preprocessing import label_binarize
from sklearn.multiclass import OneVsRestClassifier  #一對其余（每次將一個類作為正類，剩下的類作為負類）
 
from sklearn.cross_validation import train_test_split  #適用於anaconda 3.6及以前版本
#from sklearn.model_selection import train_test_split#適用於anaconda 3.7
 
#以iris數據為例，畫出P-R曲線
iris = datasets.load_iris()
X = iris.data    #150*4
y = iris.target  #150*1
 
# 標簽二值化,將三個類轉為001, 010, 100的格式.因為這是個多類分類問題，后面將要采用
#OneVsRestClassifier策略轉為二類分類問題
y = label_binarize(y, classes=[0, 1, 2])    #將150*1轉化成150*3
n_classes = y.shape[1]                      #列的個數，等於3
print (y)
 
# 增加了800維的噪聲特征
random_state = np.random.RandomState(0)
n_samples, n_features = X.shape
 
X = np.c_[X, random_state.randn(n_samples, 200 * n_features)]   #行不變，只增加了列，150*804
 
# 訓練集和測試集拆分，比例為0.5
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.5, random_state=random_state) #隨機數，填0或不填，每次都會不一樣
 
# 一對其余，轉換成兩類，構建新的分類器
classifier = OneVsRestClassifier(svm.SVC(kernel='linear', probability=True, random_state=random_state))
#訓練集送給fit函數進行擬合訓練，訓練完后將測試集的樣本特征注入，得到測試集中每個樣本預測的分數
y_score = classifier.fit(X_train, y_train).decision_function(X_test)
 
# Compute Precision-Recall and plot curve  
#下面的下划線是返回的閾值。作為一個名稱：此時“_”作為臨時性的名稱使用。
#表示分配了一個特定的名稱，但是並不會在后面再次用到該名稱。
precision = dict()
recall = dict()
average_precision = dict()
for i in range(n_classes):
    #對於每一類，計算精確率和召回率的序列（:表示所有行，i表示第i列）
    precision[i], recall[i], _ = precision_recall_curve(y_test[:, i],  y_score[:, i]) 
    average_precision[i] = average_precision_score(y_test[:, i], y_score[:, i])#切片，第i個類的分類結果性能
 
# Compute micro-average curve and area. ravel()將多維數組降為一維
precision["micro"], recall["micro"], _ = precision_recall_curve(y_test.ravel(),  y_score.ravel())
average_precision["micro"] = average_precision_score(y_test, y_score, average="micro") #This score corresponds to the area under the precision-recall curve.
 
# Plot Precision-Recall curve for each class
plt.clf()#clf 函數用於清除當前圖像窗口
plt.plot(recall["micro"], precision["micro"],
         label='micro-average Precision-recall curve (area = {0:0.2f})'.format(average_precision["micro"]))
for i in range(n_classes):
    plt.plot(recall[i], precision[i],
             label='Precision-recall curve of class {0} (area = {1:0.2f})'.format(i, average_precision[i]))
 
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05]) #xlim、ylim：分別設置X、Y軸的顯示范圍。
plt.xlabel('Recall', fontsize=16)
plt.ylabel('Precision',fontsize=16)
plt.title('Extension of Precision-Recall curve to multi-class',fontsize=16)
plt.legend(loc="lower right")#legend 是用於設置圖例的函數
plt.show()

運行結果如下：