Sounds like homework. Anyway, lookup rank correlations (Spearman,
Kendall, etc.) and correpsonding tests.
Uwe Ligges
On 14.04.2011 17:00, Baoqiang Cao wrote:> Hi,
>
> I have to one set of inputs and their observed true values of each
> input. Now I have a model takes the input and predict a value. I only
> consider the ranks based on either the observed true values or the
> predicted values. My question is how do I compare this two rank in R?
> That is, how close the rank from prediction to the rank from true
> observations?
>
> x_1: y_1, p_1,
> x_2: y_2, p_2,
> ...
> x_m: y_m, p_m,
>
> R_y is the ranking based on {y_1, y_2,..,y_m}. R_p is the ranking from
> {p_1, p_2, ..., p_m}. How do I know in R how good/bad R_p is given
> R_y?
>
> I searched r-help and got some clue of using wilcox.test but still a
> bit confused of how to compare the two ranking against the true
> ranking. Any advice will be highly appreciated. I apologize if this is
> a too statistics question.
>
> Thanks in advance.
> Baoqiang
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.