I have time series data and I'm using the cesium package to extract several features.

Now the thing is, the number of extracted features is pretty significant. I have tried various visualizations to make sense of what features classify my labels the best but it is too cumbersome for so many features.

Are there any algorithms or packages available that I can use for feature selection?

  • $\begingroup$ Have you tried PCA or TSNE? these are nothing but dimensionality reduction techniques, so my question boils down to have you tried any Dimensionality Reduction Technique? $\endgroup$
    – Kaustubh
    Oct 24, 2018 at 11:55
  • $\begingroup$ @Kaustubh I'm trying to implement various approaches mentioned here $\endgroup$ Oct 24, 2018 at 12:26
  • $\begingroup$ Ok, Do you have a strategy/rationale for using or not using any of the algorithms mentioned on that page? $\endgroup$
    – Kaustubh
    Oct 25, 2018 at 11:24


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.