pyanomaly 0.0.2.3

Time series anomaly detection in Python

pyanomaly

Conjunto de algoritmos para detectar anomalias em Series Temporais.

Instalação

pip install pyanomaly

Como usar

Iremos realizar os testes no dataset contendo temperaturas diarias da cidade de Melbourne.

dataset: https://raw.githubusercontent.com/jbrownlee/Datasets/master/daily-min-temperatures.csv

# data
import numpy as np
import pandas as pd
# plot
import matplotlib.pyplot as plt
import seaborn as sns; sns.set()

df = pd.read_csv('./dados/daily-min-temperatures.csv', parse_dates=['Date'])
df.set_index('Date', inplace=True)

print(df.head(5).T)

Date  1981-01-01  1981-01-02  1981-01-03  1981-01-04  1981-01-05
Temp        20.7        17.9        18.8        14.6        15.8

df.plot(figsize=(8, 4));

Mad

mad = MAD()
mad.fit(df['Temp'])
outliers = mad.fit_predict(df['Temp'])

outliers.head()

Date
1981-01-15    25.0
1981-01-18    24.8
1981-02-09    25.0
1982-01-17    24.0
1982-01-20    25.2
Name: Temp, dtype: float64

fig, ax = plt.subplots(1, 1, figsize=(12, 6))

sns.lineplot(x=df.index, y=df['Temp'], ax=ax)
sns.scatterplot(x=outliers.index, y=outliers, 
                color='r', ax=ax)

plt.title('Zscore Robusto', fontsize='large');

Tukey

tu = Tukey()

tu.fit(df['Temp'])
outliers = tu.predict(df['Temp'])

outliers.head()

Date
1981-01-15    25.0
1981-01-18    24.8
1981-02-09    25.0
1982-01-17    24.0
1982-01-20    25.2
Name: Temp, dtype: float64

fig, ax = plt.subplots(1, 1, figsize=(12, 6))

sns.lineplot(x=df.index, y=df['Temp'], ax=ax)
sns.scatterplot(x=outliers.index, y=outliers, 
                color='r', ax=ax)

plt.title('Tukey Method', fontsize='large');

Twitter - S-MAD

outliers = twitter(df['Temp'], period=12)
outliers.head()

Date
1981-01-15    25.0
1981-01-18    24.8
1981-02-09    25.0
1982-01-20    25.2
1982-02-15    26.3
Name: Temp, dtype: float64

fig, ax = plt.subplots(1, 1, figsize=(12, 6))

sns.lineplot(x=df.index, y=df['Temp'], ax=ax)
sns.scatterplot(x=outliers.index, y=outliers, 
                color='r', ax=ax)

plt.title('Tukey Method', fontsize='large');