I would ilke to extract features from time domain signals. However, each trial has different number of sampling points. Except from statistical features How can I extract equal number of features from these signals? Your suggestions will be appreciated.
I suggest in my opinion to use DL or ML , it's very powerful for extraction the feature , for example that DL with architecture RNN can you give better results of your problems.
I hope tha be helpful. @ Onder Aydemir
Dear Onder,
As far as I understood, your problem is nonuniform sampling. If so, you can use Nyquist – Shannon sampling theorem. I think your case is implemented in the general theory for non-baseband and nonuniform samples which was developed in 1967 by Henry Landau. This theory was later developed by other researchers and their works are available in the literature databases.
Below there is a link to Landau paper.
Landau, H. J. (1967). "Necessary density conditions for sampling and interpolation of certain entire functions." Acta Math. 117 (1): 37–52. doi: 10.1007 / BF02395039. https://projecteuclid.org/download/pdf_1/euclid.acta/1485889494
I wish you all the best
Edward
I suggest in my opinion to use DL or ML , it's very powerful for extraction the feature , for example that DL with architecture RNN can you give better results of your problems.
I hope tha be helpful. @ Onder Aydemir
If you use different number of points for feature extraction, then your obtained results should be different. But, if you chose same number of points with different index (in this situation, you may use any methods which can help you for using different index), then your results are again different from each other.
Other way you can use deep learning layer features, for example softmax layer, etc, in this case DL extract meaningful features, and so you can use unnecessary features with any methods ( canonical correlation analysis, etc), I hope your suggestion is this?
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts