- Why do people use Normal, log-normal, Weibull distributions when they do not know the exact physical distribution?
- What is the problem of using empirical distributions? What are the advantages?
- If you must choose an empirical distribution, what should be your criteria? (Nos. of parameters, physical principles, etc.)
- Why does everyone suggest the use of CDF for empirical data-fitting, rather than PDF? (Obviously one can go from one function to the other)
- There are all sorts of distribution functions (e.g. survivability function) ? If everything is related to everything else, why do we need so many?
- How would you determine the BFRW failure rates? Mean Hazard rate?
Researchers should cite this work as follows:
EE 226, Purdue University, West Lafayette, IN