我想为我在文本分类中使用的三种算法绘制一个精确率-召回率曲线。我还是个初学者，能有人告诉我如何在现有代码中添加这个功能吗？

nb_classifier = MultinomialNB()svm_classifier = LinearSVC()lr_classifier = LogisticRegression(multi_class="ovr")X_train, X_test, y_train, y_test = model_selection.train_test_split(df_train.data, df_train.label, test_size=0.2 , stratify = df_train['label'])vect = CountVectorizer(stop_words='english', max_features=10000,                       token_pattern=r'[a-zA-Z]{3,}' , ngram_range=(1,2))X_train_dtm = vect.fit_transform(X_train)X_test_dtm = vect.transform(X_test)nb_classifier.fit(X_train_dtm, y_train)svm_classifier.fit(X_train_dtm, y_train)lr_classifier.fit(X_train_dtm, y_train)nb_predictions = nb_classifier.predict(X_test_dtm)svm_predictions = svm_classifier.predict(X_test_dtm)lr_predictions = lr_classifier.predict(X_test_dtm)

回答：

您可以使用sklearn.metrics中的plot_precision_recall_curve函数来绘制这些方法的精确率-召回率曲线，如下所示：

nb_classifier = MultinomialNB()svm_classifier = LinearSVC()lr_classifier = LogisticRegression(multi_class="ovr")X_train, X_test, y_train, y_test = model_selection.train_test_split(df_train.data, df_train.label, test_size=0.2 , stratify = df_train['label'])vect = CountVectorizer(stop_words='english', max_features=10000,                       token_pattern=r'[a-zA-Z]{3,}' , ngram_range=(1,2))X_train_dtm = vect.fit_transform(X_train)X_test_dtm = vect.transform(X_test)nb_classifier.fit(X_train_dtm, y_train)svm_classifier.fit(X_train_dtm, y_train)lr_classifier.fit(X_train_dtm, y_train)nb_predictions = nb_classifier.predict(X_test_dtm)svm_predictions = svm_classifier.predict(X_test_dtm)lr_predictions = lr_classifier.predict(X_test_dtm)#plot Precision-Recall curve and display average precision-recall scorefrom sklearn.metrics import precision_recall_curvefrom sklearn.metrics import plot_precision_recall_curveimport matplotlib.pyplot as pltfrom sklearn.metrics import average_precision_scoredisp = plot_precision_recall_curve(svm_classifier, X_test_dtm, y_test) #display Precision-Recall curve for svm_classifieraverage_precision = average_precision_score(y_test, svm_predictions)print('Average precision-recall score for svm_classifier: {0:0.2f}'.format(      average_precision))disp = plot_precision_recall_curve(nb_classifier, X_test_dtm, y_test) #display Precision-Recall curve for nb_classifieraverage_precision = average_precision_score(y_test, nb_predictions)print('Average precision-recall score for nb_classifier: {0:0.2f}'.format(      average_precision))disp = plot_precision_recall_curve(lr_classifier, X_test_dtm, y_test) #display Precision-Recall curve for nb_classifieraverage_precision = average_precision_score(y_test, lr_predictions)print('Average precision-recall score for lr_classifier: {0:0.2f}'.format(      average_precision))