0
$\begingroup$

enter image description hereThis time series contains some time frame which each of them are 8K (frequencies)*151 (time samples) in 0.5 sec [overall 1.2288 millions samples per half a second)

I need to find anomalous based on different rows (frequencies) Report the rows (frequencies) which are anomalous? (an unsupervised learning method) Do you have an idea to which statistical parameter is more useful for it? mean max min median var or any parameters of these 151 sampling? Which parameter I should use? (I show one sample (151 sample per frequency) from 8k data)

$\endgroup$
5
  • $\begingroup$ The answer is data dependent and you have not shown any data. $\endgroup$
    – Jon Nordby
    Aug 23 at 13:54
  • $\begingroup$ What kind of data is this? Based on your description it sounds like Time-Frequency data (spectrogram), like the results of a STFT or simiar? Is it audio or some other domain? $\endgroup$
    – Jon Nordby
    Aug 23 at 13:55
  • $\begingroup$ Yes, it is FFT, time frequency, also I add one of 8k sampling frequency. (151 samples per each frequency) $\endgroup$ Aug 24 at 5:57
  • $\begingroup$ If this is indeed time-frequency data, then you should plot it as a spectrogram $\endgroup$
    – Jon Nordby
    Oct 13 at 20:07
  • $\begingroup$ In the plot that you have shown, what does the X and Y axis represent? $\endgroup$
    – Jon Nordby
    Oct 13 at 20:08
0
$\begingroup$

Statistical control processes and CUSUM in particular can be useful in identifying outlying values or changes in time series data.

$\endgroup$
1
  • $\begingroup$ I used cusum. What about statistical control process you mean? $\endgroup$ Aug 24 at 3:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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