Tsfresh – Automatic extraction of relevant features from time series

by restapi • 10 years ago

167 8 10 years ago

ash9r• 10 years ago

sure but would they be relevant, good features?

BWStearns• 10 years ago

Given the first pass seems often to be picking some standardish stuff to look for it sounds like this could be useful even with minimally "smart" feature selection.

_lm_• 10 years ago

To avoid extracting irrelevant features, the TSFRESH package has a built-in filtering procedure. This filtering procedure evaluates the explaining power and importance of each characteristic for the regression or classification tasks at hand.
It is based on the well developed theory of hypothesis testing and uses a multiple test procedure. As a result the filtering process mathematically controls the percentage of irrelevant extracted features.

Here's the paper on this: https://arxiv.org/abs/1610.07717

It seems that the relevance of the features is somewhat tunable based on the p-value you choose for the statistical tests. (Every feature selection algorithm I can think of has some tunable parameter, although the information theoretic ones just depend on the length of features you're willing to consider.)

MaxBenChrist2• 10 years ago

The individual feature significance tests do not have any parameter, they just generate the p-values.

The only parameter that one can tune is the overall percentage of irrelevant extracted features. That is the expected FDR of the Benjamini yakutieli procedure.

madengr• 10 years ago

Would be nice in an oscilloscope, though many have a good set of waveform measurements.

placebo• 10 years ago

I've implemented a peaks/troughs feature extraction in Javascript a few years ago as a basis for some larger analysis project. It works at O(n) of the number of data points. You can play with a demo here:

http://nocurve.com/2014/01/12/finding-peaks-and-troughs-in-a...

zump• 10 years ago

Why not use an RNN?

MaxBenChrist2• 10 years ago

what for? for the feature extraction part?

Tsfresh – Automatic extraction of relevant features from time series

Related Stories