# Selection of transforms, windows, wavelets for current signal analysis

I am new to this topic but I already read about the propperties of different signal analysis methods. I am stuck with the obvious advantages of FFT, STFT and Wavelet transforms but can't find good parameters for them.
I have a current signal from a motor that should be transformed to give me characteristic features. It will beanalysed with a machine learning algorithm. Its envelopecurve can be represented as a test parabola of the form:
$y= -0.4\cdot(x-4)^2+3$ with $0\le x\le8$ and $y\ge 0$
Right now I try different transform algorithms (FFT, STFT, PSD, DFT, CWT) with the test parabolas and different parameters to see which transform could give me hints about charakteristics of the signal in Python. I want to detect the position and magnitude of dents in the signal to use them as features.

Do you have any suggestions for resources to make a good choice of the transform algorithms, windows, wavelets in general?

• what do you mean by dents? their position and magnitude? what do you want to characterize exactly? Mar 13 '18 at 9:20