Talk:Basic statistics: mean, median, average, standard deviation, z-scores, and p-value
From ControlsWiki
- It would be helpful to see a proof of the z-score to p-value transform shown graphically at the end of the p-value section. What if the actual integration were shown and the results calculated for each case?
- this is done in the example, but might be neat to see earlier..
- a more detailed discussion of the implications of a p-value of <0.05 (maybe a set of numerical experiments) would be helpful.
- For the p-value case, what about if there are 3 hypotheses vs. just the H1 and H0? This multiple comparison case can be very tricky in a frequentist p-value analysis, but is worth at least mentioning.
- How should this article be placed in context of random distributions in general? Later we introduce continuous and discrete distributions, but we use them here...
[edit]
EWA
Should discuss (exponential) weighted averaging more in the sense of historical filtering, VERY useful for accommodating noise in software. For example, new "input" = WT1*previous input (average) + WT2*latest input. Typically 80/20 or so depending on required response.--Billymac00 14:40, 19 May 2008 (EDT)
