adadmire 0.0.8

Functions for detecting anomalies in tabular datasets using Mixed Graphical Models.

adadmire

Functions for detecting anomalies in tabular datasets using Mixed Graphical Models.

Installation

Enter the following commands in a shell like bash, zsh or powershell:

pip install -U adadmire
pip install git+https://github.com/yhat/ggplot.git

Usage

⚠️Attention: this section is currently in draft mode, i.e. the listed examples are not yet working and must be updated first.

Example 1

from adadmire import loo_cv_cor, get_threshold_continuous, get_threshold_discrete
import numpy as np

X = np.load('C:/Users/l_buc/paper_mgm/val_wolfram/data/scaled.npy')
D = np.load('C:/Users/l_buc/paper_mgm/val_wolfram/data/bm.npy')
levels = np.load('C:/Users/l_buc/paper_mgm/val_wolfram/data/levels_bm.npy')
lam_zero = np.sqrt(np.log(X.shape[1] + D.shape[1]/2)/X.shape[0])
lam_seq = np.array([-2.0,-2.25])
lam = [pow(2, x) for x in lam_seq]
lam = np.array(lam)
lam = lam_zero * lam
prob_hat, B_m, lam_opt,  x_hat_cor_xp, d_hat_cor_xp = loo_cv_cor(X,D,levels,lam)
X_cor, threshold, n_ano,  ano_index = get_threshold_continuous(X, x_hat_cor_xp, B_m)
n_ano, threshold, pos = get_threshold_discrete(D, levels, d_hat_cor_xp)

Example 2

import adadmire
import sklearn.datasets

# load diabetes dataset from sklearn
diab = sklearn.datasets.load_diabetes()
# 442x4 matrix with scaled features: age, sex, bmi, blood pressure
X = diab.data[:, 1:4]
y = diab.target
# Lets introduce some faulty entries
X[100, 1] = 0.8
X[200, 2] = 0.7
X[300, 3] = 0.1
X[400, 4] = 0.2
# Lets detect (and correct) them using adadmire
ca = adadmire.detect_anomalies(X)
print(ca)
X_corrected = adadmire.correct_anomalies(X)

Contribute

In case you have questions, feature requests or find any bugs in adadmire, please create a corresponding issue at gitlab.spang-lab.de/bul38390/admire/issues.

In case you want to write code for this package, see Contribute for details.