基于无标签机器学习的异常检测

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import random
import os
import time
import math
import iForest

# 生成随机数据
def generate_data(n, n_features, n_anomalies):
    data = np.random.randn(n, n_features)
    for i in range(n_anomalies):
        data[i] = [random.uniform(-10, 10), random.uniform(-10, 10)]
    return data

# 使用Isolation Forest进行异常检测
def detect_anomalies(data, n_estimators=100):
    forest = iForest.IsolationForest(n_estimators=n_estimators, random_state=42)
    forest.fit(data)
    anomalies = forest.predict(data)
    return anomalies

# 示例数据
data = generate_data(1000, 2, 10)
anomalies = detect_anomalies(data)

# 绘图
plt.figure(figsize=(10, 6))
sns.scatterplot(x=data[:, 0], y=data[:, 1], hue=anomalies, palette="coolwarm", legend='full')
plt.title('Anomaly Detection using Isolation Forest')
plt.show()