I am working on a machine learning problem, I would like to see the network performance in classifying an input data set after injecting a normal Gaussian noise to the clean input data set. I used MATLAB and python to generate the noise but not sure how to convert to decibel(DB).
Additive white Gaussian noise (AWGN) is a basic noise model used in Information theory to mimic the effect of many random processes that occur in nature. The modifiers denote specific characteristics:
- 'Additive' because it is added to any noise that might be intrinsic to the information system.
- 'White' refers to idea that it has uniform power across the frequency band for the information system. It is an analogy to the color white which has uniform emissions at all frequencies in the visible spectrum. (This means if you take the Fourier Transform of the noise, and plot the power spectrum (power vs frequency), you will obtain a uniform distribution)
- 'Gaussian' because it has a normal distribution in the time domain with an average time domain value of zero. (This means if you construct the PDF of the AWGN, it will follow a gaussian distribution)
For example,
Below is the distribution of a WGN generated using "wgn" command in MATLAB for 10000 elements. 📷Now is the power spectral density estimate of the same noise using "pwelch" command.
📷PSD is uniform, in the sense that the power of all frequencies are within a particular range.
White noise does not necessarily mean gaussian. Similarly Gaussian does not mean it is white noise.
Hope this helps.
- Badges
- Science topic