Talk:Process capability index: Difference between revisions

Revision as of 23:17, 5 November 2016

Latest revision as of 11:53, 8 February 2024

(5 intermediate revisions by 3 users not shown)

Line 1:

{{WikiProject Businessbanner shell|class=C|importance=Low}}

{{WikiProject Statistics|class=CBusiness|importance=Low}}

{{WikiProject Statistics|importance=Low}}

}}

{{User:ClueBot III/ArchiveThis|archiveprefix=Talk:Process capability index/Archives/|format=Y|age=26297|index=yes|archivebox=yes|box-advert=yes}}

---

==Untitled==

I propose to use the example of a runway instead. Everybody can imagen that you would like the pilot to land at the middle of the runway each time.

-gr, JVP <span style="font-size: smaller;" class="autosigned">—Preceding [[Wikipedia:Signatures|unsigned]] comment added by [[Special:Contributions/145.36.63.11|145.36.63.11]] ([[User talk:145.36.63.11|talk]]) 14:45, 20 January 2011 (UTC)</span><!-- Template:UnsignedIP --> <!--Autosigned by SineBot-->

I corrected the notation for "variability of the process" to standard deviation to avoid confusion between sigma and sigma squared, variance/variability, and standard deviation.

Hyoga 11/26/2007

----

There is no point in listing an equation if you are not going to define what the variables mean. Namely, what on earth is do USL and LSL stand for? Even sigma should be defined for those who do not know what they are looking at.

:Fixed. --mabahj 11:00, 12 July 2006 (UTC)

USP - Upper Specification Limit

LSL - Lower specification Limit

These are related to statistical process control (SPC).

-- Manaf 28 June 2007 (EST)

== Table for sigma levels ==

Either this table, or the table in the wikipedia article with the following link is wrong:

http://en.wikipedia.org/wiki/Six_Sigma#Sigma_levels

I am not an expert, so please someone revise it.

--[[Special:Contributions/59.66.122.148|59.66.122.148]] ([[User talk:59.66.122.148|talk]]) 13:08, 9 January 2014 (UTC)

:No, they're correct. For process capability to be valid, the process must be stable, so yields of Φ(sigma level) - Φ(-sigma level) as given in the table (in the article coupled with this talk page) are correct. As the [[Six Sigma]] article indicates, the yields presented there represent the worst case deterioration of 1.5σ before a [[control chart]] is expected to catch the shift or drift, so yields of Φ(sigma level - 1.5) - Φ(-sigma level - 1.5) (for a shift or drift upward; a shift or drift downward is equivalent—you can do the math) are correct in that article. No action required. -- [[User:DanielPenfield|DanielPenfield]] ([[User talk:DanielPenfield|talk]]) 10:02, 27 January 2014 (UTC)

Example shown has a std. deviation that seems to not match chart data. Mu hat estimate (98.94) is correct, but estimated chart data std. deviation is ~0.869 (unbiased) or ~0.842 (std. population deviation), not 1.03, as stated.

..unless I am missing something or a different method is used for this? Any comments/clarification very welcomed and appreciated.

[[Special:Contributions/165.225.34.90|165.225.34.90]] ([[User talk:165.225.34.90|talk]]) 23:09, 5 November 2016 (UTC)