def __init__(self, master, x_train, y_train, x_test, y_test, evaluator):

Tk.Frame.__init__(self, master)

# evaluator.load_data(x_train, y_train, x_test, y_test)

# evaluator.train()

x_test_r = evaluator.reduce(x_test) # 特征降维

frame_test = Tk.Frame(self)

frame_test.pack(fill='x', expand=1, padx=15, pady=15)

figure_test = Figure(figsize=(4, 4), dpi=100)

subplot_test = figure_test.add_subplot(111)

subplot_test.set_title('Breast Cancer Testing')

figure_test.tight_layout()

h = .02 # step size in the mesh

x1_min, x1_max = x_test_r[:, 0].min() - 1, x_test_r[:, 0].max() + 1

x2_min, x2_max = x_test_r[:, 1].min() - 1, x_test_r[:, 1].max() + 1

xx1, xx2 = np.meshgrid(np.arange(x1_min, x1_max, h), np.arange(x2_min, x2_max, h))

yy = evaluator.clf.predict(np.c_[xx1.ravel(), xx2.ravel()])

yy = yy.reshape(xx1.shape)

subplot_test.contourf(xx1, xx2, yy, cmap=plt.cm.get_cmap("Paired"), alpha=0.8)

subplot_test.scatter(x_test_r[:, 0], x_test_r[:, 1], c=y_test, cmap=plt.cm.get_cmap("Paired"))

self.attach_figure(figure_test, frame_test)

# 第5.1页 测试性能指标 precision recall f_value

y_pred = evaluator.pipeline.predict(x_test)

frame_matrix = Tk.Frame(self)

frame_matrix.pack(side=Tk.LEFT, fill='x', expand=1, padx=15, pady=15)

figure_matrix = Figure(figsize=(4, 4), dpi=100)

subplot_matrix = figure_matrix.add_subplot(111)

confmat = confusion_matrix(y_true=y_test, y_pred=y_pred)

subplot_matrix.matshow(confmat, cmap=plt.cm.get_cmap("Blues"), alpha=0.3)

for i in range(confmat.shape[0]):

for j in range(confmat.shape[1]):

subplot_matrix.text(x=j, y=i, s=confmat[i, j], va='center', ha='center')

subplot_matrix.set_xlabel('predicted label')

subplot_matrix.set_ylabel('true label')

self.attach_figure(figure_matrix, frame_matrix)

frame_result = Tk.Frame(self)

frame_result.pack(side=Tk.LEFT, fill='x', expand=1, padx=15, pady=15)

Tk.Label(frame_result, text="Accuracy: ").grid(row=0, column=0, sticky=Tk.W)

Tk.Label(frame_result, text=str(evaluator.pipeline.score(x_test, y_test))).grid(row=0, column=1,

sticky=Tk.W)

Tk.Label(frame_result, text="Precision: ").grid(row=1, column=0, sticky=Tk.W)

Tk.Label(frame_result, text=str(precision_score(y_true=y_test, y_pred=y_pred))).grid(row=1, column=1,

sticky=Tk.W)

Tk.Label(frame_result, text="Recall: ").grid(row=2, column=0, sticky=Tk.W)

Tk.Label(frame_result, text=str(recall_score(y_true=y_test, y_pred=y_pred))).grid(row=2, column=1,

sticky=Tk.W)

Tk.Label(frame_result, text="F-value: ").grid(row=3, column=0, sticky=Tk.W)

Tk.Label(frame_result, text=str(f1_score(y_true=y_test, y_pred=y_pred))).grid(row=3, column=1, sticky=Tk.W)

@staticmethod

def attach_figure(figure, frame):

canvas = FigureCanvasTkAgg(figure, master=frame) # 内嵌散点图到UI

canvas.show()

canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

toolbar = NavigationToolbar2TkAgg(canvas, frame) # 内嵌散点图工具栏到UI

toolbar.update()