Anomaly detection using wavelet transforms

In a previous note, we discussed how to detect anomalies using linear filters. Thanks to the comments we received, in this note we show how to solve the same problem, but now using wavelet transforms.

In this case we are using PyWavelets, which is a free open source library in Python, available here:

https://pywavelets.readthedocs.io/en/latest/index.html

In fact, in this case we only need to compare the details decomposition to a threshold, to be able to detect the anomaly. You can simply refer to the figure showed above, in green color. You can run the code, available here:

Do you know a better approach to solve this problem? Do you have a counterexample in which this method does not work? Do you have any general comment about this note?

I would be happy to receive your comments to: [email protected]

At MYWAI we promote agile, explainable, reliable and affordable ML at the edge.