0
$\begingroup$

I am trying to extract some features of the accelerometer sensor and during the feature extraction each 5 second segment in the array is windowed and features are generated from these windows. Features extracted include: • Mean of the magnitude data • Maximum value. • Minimum value. • Standard Deviation value for each axis (x, y, and z). • Average Standard Deviation over 3 axes. • Root mean squared acceleration (RMS). • The sum of height of frequency component below 5 Hz. • Number of peaks in spectrum.

I am not sure if mixing all these is adding redundant values and affect the final results.

Thanks

$\endgroup$
0
$\begingroup$

Whether or not these features are redundant can only be concluded regarding your target variable that you want to deduce. Mathematically speaking, these features are not redundant, since you can not directly compute one from the other without additional information.

Thus each feature adds additional information.

For redundancy check in general, you can compute the mutual information or - often used - check for correlation.

$\endgroup$

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.