Signal processing tools are mostly based on the linear model, Gaussian statistics and stationary assumptions, While the world is nonstationary, non Gaussian and nonlinear. Why have the linear model and the Gaussian assumptions been so effective approximations?
When a physical variable Y has a unknown but definite dependence on a physical variable X , then Y depending on X as a linear relation is the simplest one to start with.
When a physical variable Y is expected to be completely randomly varying with X , about a value Y0 for the variable X0 , then the function for (Y-Y0) as a Gaussian function in terms of (X-X0) is the most likely to be the best function choice to start with .
Dr. Chander Mohan Singal
When a physical variable Y has a unknown but definite dependence on a physical variable X , then Y depending on X as a linear relation is the simplest one to start with.
When a physical variable Y is expected to be completely randomly varying with X , about a value Y0 for the variable X0 , then the function for (Y-Y0) as a Gaussian function in terms of (X-X0) is the most likely to be the best function choice to start with .
Dr. Chander Mohan Singal
- Badges
- Science topic