[R] why my R^2 is so small while there do seem to be a dependency there?

[Bert Gunter]

The model given, ~0 +factor, omitted an intercept. R^2 is meaningless
without an intercept.

Further discussion should go to a statistics list like
stats.stackexchange.com, as these are statistics, not R, issues.

Cheers,
Bert

On Sun, Aug 4, 2013 at 1:41 PM, Rui Barradas <ruipbarradas at sapo.pt> wrote:
> Hello,
>
> Hoe did you get 0.002? Can you ?dput your data?
>
> d <- read.table(text = "
> factor  observation
> -0.003       -2
> -0.002       -2
> -0.001       -1
> 0.000         1
> 0.001         0
> 0.002         1
> 0.003         2
> ", header = TRUE)
>
> fit <- lm(observation ~ 0 + factor, data = d)
> summary(fit)  # R2 is 0.8595, not 0.002
>
> Hope this helps,
>
> Rui Barradas
>
> Em 04-08-2013 18:19, CHEN, Cheng escreveu:
>>
>> Hi gurus!
>>
>> What I need to do is to find a model, which can predict what the *
>> observation* should look like given a *factor* input.
>>
>> i am doing a simple linear fit in R:
>>
>> lm(observation~0+factor, data=d), the R^2 is 0.002, which is really small.
>>
>> however, when I do a 'SELECT AVG observation by 0.001 BRACKET factor',
>> there result is something like:
>>
>> *factor* | *average observersion*
>>
>> -0.003       -2
>>
>> -0.002       -2
>>
>> -0.001       -1
>>
>> 0.000         1
>>
>> 0.001         0
>>
>> 0.002         1
>> 0.003         2
>>
>>
>> from a user perspective, i definitely see a pattern here, but somehow this
>> pattern is not captured by a linear model. Is my understanding correct?
>>
>> so R gurus, which model do you suggest me to try for such data?
>>
>> Thanks!
>>
>>
>
> ______________________________________________
> 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.



-- 

Bert Gunter
Genentech Nonclinical Biostatistics

Internal Contact Info:
Phone: 467-7374
Website:
http://pharmadevelopment.roche.com/index/pdb/pdb-functional-groups/pdb-biostatistics/pdb-ncb-home.htm