四种经典机器学习算法测试——Bayes, SVM, 决策树, K-Means

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

import cv2
import numpy as np
class DataLoader(object):
  def __init__(self, path, label, relabel='false'):
​    self.path = path
​    self.label = label
​    if relabel:
​      self.label = [0,1,2,3,4,5,6,7,8,9]
  def data(self):
​    x = []
​    y = []
​    for i in range(0,10):
​      for f in range(0,100):
​        l = str(100*i+f)
​        # 读取图片数据
​        images = cv2.imread("{}/{}.jpg".format(self.path,l))
​        # 统一shape
​        # image = cv2.resize(images, (256,256), interpolation=cv2.INTER_CUBIC)
​        # 灰度直方图
​        # hist = cv2.calcHist([images], [0, 1, 2], None, [256,256,256], [0.0, 256.0, 0.0, 256.0, 0.0, 256.0])

​        hist = cv2.calcHist([images], [0, 1], None, [256, 256], [0.0, 255.0, 0.0, 255.0])
​        x.append(((hist / 255).flatten()))
​        y.append(self.label[i])
​    x = np.array(x)
​    y = np.array(y)
​    return x,y

1
2

X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.3, random_state=1)	
#700张训练集300张测试集

1
2
3
4
5
6
7
8

#https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.GaussianNB.html#sklearn.naive_bayes.GaussianNB
def clf_byss():
    time1 = datetime.datetime.now()
    clf_bysp = MultinomialNB()
    clf_bys = clf_bysp.fit(X_train,y_train)
    predict_labels = clf_bys.predict(X_test)
    time2 = datetime.datetime.now()
    return predict_labels, y_test, time2-time1

1
2
3
4
5
6
7
8

#https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC
def clf_svms():
    time1 = datetime.datetime.now()
    clf_svmp = svm.SVC(kernel='linear', C=0.01)
    clf_svm = clf_svmp.fit(X_train,y_train)
    predict_labels = clf_svm.predict(X_test)
    time2 = datetime.datetime.now()
    return predict_labels, y_test, time2-time1

1
2
3
4
5
6
7
8

#https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html?highlight=decisiontree#sklearn.tree.DecisionTreeClassifier
def clf_dtrees():
    time1 = datetime.datetime.now()
    clf_dtreep = tree.DecisionTreeClassifier(max_depth=20)
    clf_dtree = clf_dtreep.fit(X_train,y_train)
    predict_labels = clf_dtree.predict(X_test)
    time2 = datetime.datetime.now()
    return predict_labels, y_test, time2-time1

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

#https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html?highlight=kmeans#sklearn.cluster.KMeans
def clf_kms():
    Xk,Yk = dl.DataLoader(path, labels, relabel='true').data()
    X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.3, random_state=1)
    time1 = datetime.datetime.now()
    clf_kmp = KMeans(n_clusters=10,max_iter=300)
    clf_km = clf_kmp.fit(X, Y)
    predict_labels = clf_km.predict(X_test)
    time2 = datetime.datetime.now()
    return  predict_labels, yk_test, time2-time1

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

titles_options = [("Confusionmatrix_Bayes", clf_byss()),
                  ('Confusionmatrix_SVM', clf_svms()),
                  ('Confusionmatrix_Decision-Tree', clf_dtrees()),
                  ('Confusionmatrix_KMeans', clf_kms())]
for title, result in titles_options:
    predict_labels, y_test, cost = result[0], result[1], result[2]
    con_mat = confusion_matrix(y_test, predict_labels)  #混淆矩阵
    con_mat_norm = con_mat.astype('float') / con_mat.sum(axis=1)[:, np.newaxis]  # 归一化
    con_mat_norm = np.around(con_mat_norm, decimals=2)
		#测试报告
    reports = classification_report(y_test, predict_labels, target_names=labels, zero_division=1)
		accuracy_scores = accuracy_score(y_test, predict_labels)
    #https: // seaborn.pydata.org / generated / seaborn.heatmap.html
    plt.figure(figsize=(12, 12), dpi=120)#可视化
plt.text(11,-1, 'Acc: {:.2f}'.format(accuracy_scores), fontdict={'fontsize': 15,'fontweight' : 1000})
    ax = sn.heatmap(con_mat_norm, annot=True, xticklabels=labels,yticklabels=labels,cmap=plt.cm.Blues )
    ax.set_title(title, fontdict={'fontsize': 20,'fontweight' : 1000})  # 标题
    ax.set_xlabel('Predict Label', fontdict={'fontsize': 15,'fontweight' : 1000})  # x轴
    ax.set_ylabel('True Label', fontdict={'fontsize': 15,'fontweight' : 1000})  # y轴
    #结果存储
    with open('./result/OutputsH.txt','a') as f:
        f.write('{} Report:\n{}\nCost:\n{}\n\n\n'.format(title[16::], reports, cost))
    # print('{} Report:\n{}\n\tCost:\n\t{}\n\n'.format(title[16::], reports, cost))
    plt.savefig('./result/{}.jpg'.format(title[16::],time))
f.close()