[R] Fisher

[Uwe Ligges]

Ambrosini Alessandro wrote:
> Hello.
> I had a big collection of Web pages. Now I have this collection divided into
> clusters. Every page can be relevant or not. I made a table:
>           relevant     non relevant
> cluster1    1               20
> cluster2    0               15
> cluster3    3               35
>    .        .               .
>    .        .               .
>    .        .               .
> I cluster1 I have 21 Web pages, 1 relevant and 20 no.
> I want to find if relevant pages tend to stay in some clusters, and so I
> want to find if there is a dipendence relevant-cluster. The problem is that
> I have not much relevant pages for cluster. They are 1,2,3 max 5 for cluster
> and so I can't use Chi- square of Pearson.
> Tell me one thing: suppose to have
> 
>>a
> 
> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11]
> [1,]    0    0    1    0    0    0    0    0    0     1     0
> [2,]   21   20   33   17   12   18   12   10   11    10    28
> 
> In this case every column is a cluster, the first row has the relevant pages
> ...
> if I do
> 
> 
> fisher.test(a)
> 
>         Fisher's Exact Test for Count Data
> 
> data:  a
> p-value = 0.5611
> alternative hypothesis: two.sided
> 
> how can I interpret this output? How can I read the p-value? 

[This seems to be off topic on this list.]

Well, interpreting a p-value is a *very* basic task in statistics. You 
really should read a statistical textbook on inference theory, if you 
want to work with statistical tests ...


> Have I to compare it with something? 

Your \alpha ?


> In the case of perfect dependence, is p-value=1 ?

Almost 0!!!
You are testing on independency!
The null hypothese is rejected, if p < \alpha.


> Please help me, I can't use any book of statistic in this moment and so I
> cant solve the problem by myself. My work can not go on if I don't solve the
> problem.

Uwe Ligges

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._