用AutoEncoder做异常检测

#导入相关的库并生成数据
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pyod.utils.data import generate_data
contamination = 0.05 # percentage of outliers
n_train = 500       # number of training points
n_test = 500        # number of testing points
n_features = 25      # number of features
X_train, X_test, y_train, y_test = generate_data(
    n_train=n_train, 
    n_test=n_test, 
    n_features= n_features, 
    contamination=contamination, 
    random_state=123)

X_train_pd = pd.DataFrame(X_train)
X_train_pd.head()


# Plot
plt.scatter(X_train_pd[0], X_train_pd[1], c=y_train, alpha=0.8)
plt.title('Scatter plot')
plt.xlabel('x0')
plt.ylabel('x1')
plt.show()
    

from pyod.models.auto_encoder import AutoEncoder     # pyod后续版本是基于torch的
atcdr = AutoEncoder(contamination=0.05, hidden_neuron_list= [2, 2])
atcdr.fit(X_train)  #内部会自动进行标准化

# Training data  训练数据的分数 和预测值
y_train_scores = atcdr.decision_function(X_train)
y_train_pred = atcdr.predict(X_train)

# Test data  测试数据的分数和预测值
y_test_scores = atcdr.decision_function(X_test)
y_test_pred = atcdr.predict(X_test) # outlier labels (0 or 1)

# Threshold for the defined comtanimation rate  contamination 5% 对应的值；
print("The threshold for the defined contamination rate:" , atcdr.threshold_)

def count_stat(vector):
    # Because it is '0' and '1', we can run a count statistic. 
    unique, counts = np.unique(vector, return_counts=True)
    return dict(zip(unique, counts))

print("The training data:", count_stat(y_train_pred))
print("The training data:", count_stat(y_test_pred))

结果：
The threshold for the defined contamination rate: 4.2399105072015875
The training data: {0: 475, 1: 25}
The training data: {0: 475, 1: 25}

plt.figure(figsize=(6, 4), dpi=300)
import matplotlib.pyplot as plt
plt.hist(y_train_scores, bins='auto') # arguments are passed to np.histogram
plt.title("Outlier score")
plt.show()

threshold = atcdr.threshold_ # Or other value from the above histogram
threshold
4.2399105072015875