aakr (Auto Associative Kernel Regression)

aakr is a Python implementation of the Auto-Associative Kernel Regression (AAKR). The algorithm is suitable for signal reconstruction, which can further be used for condition monitoring, anomaly detection etc.

Documentation is available at https://aakr.readthedocs.io.

Installation

pip install aakr

Quickstart

Given historical normal condition X_nc examples and new observations X_obs of size n_samples x n_features, what values we expect to see in normal conditions for the new observations?

from aakr import AAKR

# Create AAKR model
aakr = AAKR()

# Fit the model with normal condition examples
aakr.fit(X_nc)

# Ask for values expected to be seen in normal conditions
X_obs_nc = aakr.transform(X_obs)

References

• Assessment of Statistical and Classification Models For Monitoring EDF’s Assets

• A modified Auto Associative Kernel Regression method for robust signal reconstruction in nuclear power plant components

---

Jesse Myrberg (jesse.myrberg@gmail.com)

Documentation

• Getting started
  • Installation
  • Quickstart
  • References
• Usage examples
  • Simple example
  • Online example

API

• aakr package
  • Module contents