0
$\begingroup$

I am experimenting with EEG dataset from this (title: "The Perils and Pitfalls of Block Design for EEG Classification Experiments") paper. They have applied various ML models on different window sizes like $440 ms$, $200 ms$ etc.

I want to know are they epochs of the data, or some hamming window. How to code them.

Improve this question
$\endgroup$
2
  • 2
    $\begingroup$ madiha Rehman welcome to Signal Processing SE. In order for someone to answer your question, they have to go through the paper and this is not something that most people either have the time or the will to do just to answer your question. I suggest you make your question more detailed describing exactly the problem as well as the context. Add as much related information as possible in the question and don't ask from people to go through papers that may even not have access to. $\endgroup$
    – ZaellixA
    Jun 10 at 16:45
  • $\begingroup$ I would suggest posing the question with some images inserted, perhaps screenshots from the IEEE paper or from another source to frame and describe your question. Off hand, an epoch is a big transition in a waveform that demarks a "before" and "after". I have never heard the term "epoching". Windowing is something we do, and sometimes windows are aligned relative to some recognized epoch, either to center the epoch mark in the center of the window or the opposite; to put the epoch boundaries outside the window. $\endgroup$
    – robert bristow-johnson
    Jun 10 at 20:33

1 Answer 1

Reset to default
0
$\begingroup$

You can check the definition of epochs in this discussion: https://dsp.stackexchange.com/a/41145/48080

Windowing means taking a short period of EEG signals as temporal features, and the size of the windows is an important factor of the model performance.[1]

The window construction depends on problems. For example, to predict the gait at time $t$, a window can be constructed by selecting the EEG signals within the period $[t-\triangle t, t]$, where $\triangle t$ is window size. Then the feature matrix has the shape of [CHANNEL_NUMBER, WINDOW_SIZE].

Hope my answer would help :)

[1] https://ieeexplore.ieee.org/document/9892435/

Improve this answer
$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.