Statistics Help @ Talk Stats Forum - Statistics
http://www.talkstats.com/
Statistics course and homework discussion. Elementary statistics.enThu, 23 Nov 2017 14:31:41 GMTvBulletin60http://www.talkstats.com/images/misc/rss.pngStatistics Help @ Talk Stats Forum - Statistics
http://www.talkstats.com/
<![CDATA[Earth mover's distance]]>
http://www.talkstats.com/showthread.php/70820-Earth-mover-s-distance?goto=newpost
Thu, 23 Nov 2017 11:58:25 GMTI am working on earth mover's distance. I am searching for an example from some reputed journal or book so that I can validate my code.Kindly,share anyvalid dataset with calculated EMD.
Thanks in advance:)
]]>StatisticsDivya divyahttp://www.talkstats.com/showthread.php/70820-Earth-mover-s-distanceHow to interpret HZ with a quantitative variable
http://www.talkstats.com/showthread.php/70818-How-to-interpret-HZ-with-a-quantitative-variable?goto=newpost
Thu, 23 Nov 2017 11:27:43 GMTHi,
I am writing to you to get some help for a very basic question. I am performing a cox model to evaluate death according to a main quantitative...Hi,

I am writing to you to get some help for a very basic question. I am performing a cox model to evaluate death according to a main quantitative variable which is 1.5 1.6 1.62 1.74 1.99 2.5 3.15 4.56 etc...I mean it is not a variable with entire numbers as 1 2 3 etc. I have a HR = 1.25. There is an increase in the risk of death but theoritically it is for one unit increase in my quantitative variable? When you work on age for example, it is easy to understand there is an increased for each additional year but in my case I do not really understand what unit increase mean? It is 1 2 3 ? or 1.1 1.2 1.3 1.4 1.5 ??

I would also like to know how it works when you have a variable both negative and positive. For example -2 -1.5 -0.9 0 0.4 1.6 2.7 ?? How can I interpret my HR. For which kind of unit increase is it?

I am really sorry for this question but I wanted a confirmation.

Thank you so much for your help.

Best regards,

Pierre M
]]>StatisticsPierre Martinhttp://www.talkstats.com/showthread.php/70818-How-to-interpret-HZ-with-a-quantitative-variableHow many tests to run to be 90% confident of results distribution
http://www.talkstats.com/showthread.php/70817-How-many-tests-to-run-to-be-90-confident-of-results-distribution?goto=newpost
Thu, 23 Nov 2017 09:10:59 GMTHi

I'm running some tests where for each individual test it will only produce 1 result from N different possible results.

The N different outcomes are not equally probable.

After each test I'm recording the results, but I'm not sure how (to calculate how) many times I need to repeat the test to be 90 or 95% confident that my results distribution is accurate.

For example, I'm rolling an unfair 12 sided die. How many times should I roll it to be 90% confident of the result distribution?

Many thanks for any help and / or pointing in the right direction.

EDIT: in my rough test so far, carrying on the dice example, it "looks" like results 1-6 come up each 1/24 times, 7-10 2/24, 11 4/24 and 12 6/24, if that makes a difference