# -*- coding: utf-8 -*-
import numpy as np
import scipy as sp
import matplotlib.pyplot as plt
from sklearn import tree
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import classification_report
from sklearn.cross_validation import train_test_split 
plt.switch_backend('agg') 

# 資料讀入 
data = []
labels = []
with open("file_fac_abe.txt") as ifile:
	for line in ifile:
		tokens = line.strip().split(',')
		#print('tokens:',tokens)
		data.append([int(tk) for tk in tokens[:-1]])
		labels.append(tokens[-1])

x = np.array(data)
labels = np.array(labels)
y = np.zeros(labels.shape)
print('x:',x)
print(len(x))

print('labels:',labels)
print(len(labels))

print('y:',y)
print(len(y))
 
#標籤轉換為0/1
y[labels=='LCS']=1
print('y:',y)
 
#拆分訓練資料與測試資料
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.2,random_state=0)

'''
print('x_train:',x_train)
print(len(x_train))
print('x_test:',x_test)
print(len(x_test))
print('y_train:',y_train)
print(len(y_train))
print('y_test:',y_test)
print(len(y_test))
'''

# 使用資訊熵作為劃分標準，對決策樹進行訓練 
clf = tree.DecisionTreeClassifier(criterion='entropy')
#print('clf:',clf)
clf.fit(x_train, y_train)
print('clf:',clf)

# 把決策樹結構寫入檔案 '''
with open("DT_fac_abe_tree.dot", 'w') as f:
	f = tree.export_graphviz(clf, out_file=f)

# 係數反映每個特徵的影響力。越大表示該特徵在分類中起到的作用越大 '''
print('clf.feature_importances_:',clf.feature_importances_)

#測試結果的列印
answer = clf.predict(x_test)

print('x_test:',x_test)
print('answer:',answer)
print('y_test:',y_test)

print('np.mean(answer==y_test):',np.mean( answer == y_test))

#準確率與召回率
precision, recall, thresholds = precision_recall_curve(y_test, clf.predict(x_test))
answer = clf.predict_proba(x_test)[:,1]
print(classification_report(y_test, answer, target_names = ['Non_LCS','LCS']))

 

檔案為“file_fac_abe.txt”，其格式為：

得到結果如下：

（train_set_model）

(test_results) :