- 手动实现基于高斯分布的Algorithm异常检测算法
"""
@Title: algorithm
@Time: 2024/2/16
@Author: Michael Jie
"""
import numpy as np
# 高斯分布
def guassian(x, mu, sigma):
return (1 / (np.sqrt(2 * np.pi) * sigma)) * np.exp(-((x - mu) ** 2) / (2 * sigma ** 2))
print("----------手动实现algorithm----------")
# 数据集
data_sets = [0.5, 4, 6, 5, 4.7, 7, 6, 3.6, 3.8]
# 数据集长度
m = len(data_sets)
# 均值
mu = sum(data_sets) / m
# 方差
sigma = 0
for data_set in data_sets:
dev = abs(data_set - mu)
sigma += dev ** 2
sigma = np.sqrt(sigma / m) # 标准差
print("均值为:{mu},方差为:{sigma}".format(mu=mu, sigma=sigma))
# 测试
epsilon = 0.01
datas = [0.1, 4, 12.5, 27, 42.3]
for data in datas:
p = guassian(data, mu, sigma)
if p >= epsilon:
print("{data}为正常数据".format(data=data))
else:
print("{data}为异常数据".format(data=data))
"""
----------手动实现algorithm----------
均值为:4.511111111111111,方差为:1.782285950647803
0.1为正常数据
4为正常数据
12.5为异常数据
27为异常数据
42.3为异常数据
"""