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

Tk.Frame.__init__(self, master)

frame_linear_param = Tk.Frame(self)

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

figure_gs = Figure(figsize=(6, 4), dpi=100)

subplot_linear_param = figure_gs.add_subplot(111)

figure_gs.tight_layout()

subplot_linear_param.set_xscale('log')

subplot_linear_param.set_ylim([0.5, 1.0])

subplot_linear_param.set_xlabel("C")

subplot_linear_param.set_ylabel("Accuracy")

subplot_linear_param.set_title("GridSearchCV on parameter C in SVM with linear-kernel")

param_range = [0.0001, 0.001, 0.01, 0.1, 1.0, 10.0, 100.0, 1000.0]

param_grid = [{'clf__C': param_range, 'clf__kernel': ['linear']},

{'clf__C': param_range, 'clf__gamma': param_range, 'clf__kernel': ['rbf']}]

gs = GridSearchCV(estimator=evaluator.pipeline, param_grid=param_grid, scoring='accuracy', cv=10,

n_jobs=-1)

gs = gs.fit(x_train, y_train)

y_true, y_pred = y_test, gs.predict(x_test)

lieanr_end = len(param_range)

subplot_linear_param.plot(map(lambda e: e[0]['clf__C'], gs.grid_scores_[0:lieanr_end]),

map(lambda e: e[1], gs.grid_scores_[0:lieanr_end]),

linewidth=1, label='svm_linear', color="blue", marker='o', markersize=5)

subplot_linear_param.grid()

self.attach_figure(figure_gs, frame_linear_param)

frame_rbf_param = Tk.Frame(self)

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

scrollbar = Tk.Scrollbar(frame_rbf_param)

scrollbar.pack(side=Tk.RIGHT, fill=Tk.Y)

text_rbf_param = Tk.Text(frame_rbf_param, width=800, wrap='none', yscrollcommand=scrollbar.set)

text_rbf_param.insert(Tk.END, "1. Best parameter: " + str(gs.best_params_) + "\n\n")

text_rbf_param.insert(Tk.END, "2. Best parameter performance on testing data.\n\n ")

text_rbf_param.insert(Tk.END, classification_report(y_true, y_pred))

text_rbf_param.insert(Tk.END, "\n\n")

text_rbf_param.insert(Tk.END, "3. All parameter searched by GridSearchCV.\n\n ")

log = []

for params, mean_score, scores in gs.grid_scores_:

log.append(

["%0.3f" % (mean_score),

"(+/-%0.03f)" % (scores.std() * 2),

params["clf__C"],

params.has_key("clf__gamma") and params["clf__gamma"] or "",

params["clf__kernel"]])

text_rbf_param.insert(Tk.END, tabulate(log, headers=["Accuracy", "SD", "C", "gamma", "type"]))

text_rbf_param.pack()

scrollbar.config(command=text_rbf_param.yview)

@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()