[R] feature not available

[Liaw, Andy]

From: Stephen Choularton
> 
> Hi
>  
> I am working with this data:
>  
>  
> my data summary is:
>  
>  
> > summary(spi)
>       open           high           low           close       
>    volume
> 
>  Min.   :4315   Min.   :4365   Min.   :4301   Min.   :4352   Min.   :
> 0  
>  1st Qu.:4480   1st Qu.:4497   1st Qu.:4458   1st Qu.:4475   1st
> Qu.:11135  
>  Median :4609   Median :4631   Median :4594   Median :4614   Median
> :14439  
>  Mean   :4620   Mean   :4640   Mean   :4599   Mean   :4620   Mean
> :16590  
>  3rd Qu.:4773   3rd Qu.:4796   3rd Qu.:4753   3rd Qu.:4766   3rd
> Qu.:18294  
>  Max.   :4944   Max.   :4954   Max.   :4912   Max.   :4937   Max.
> :73559  
>   openInterest      direction           volatility    
> volumeXdirection  
>  Min.   :     0   Min.   :-62.00000   Min.   : 0.00   Min.   
> :-2795242  
>  1st Qu.:184685   1st Qu.:-19.25000   1st Qu.:30.00   1st 
> Qu.: -248740  
>  Median :193233   Median : -1.50000   Median :38.50   Median 
> :  -15905  
>  Mean   :188825   Mean   :  0.01563   Mean   :41.58   Mean   
> :    6275  
>  3rd Qu.:199800   3rd Qu.: 17.00000   3rd Qu.:50.00   3rd 
> Qu.:  206325  
>  Max.   :236759   Max.   : 74.00000   Max.   :94.00   Max.   
> : 2024470  
>  volatilityXdirection     upDown          nextDay      
>  Min.   :      0      Min.   :0.0000   Min.   :0.0000  
>  1st Qu.: 362816      1st Qu.:0.0000   1st Qu.:0.0000  
>  Median : 540187      Median :0.0000   Median :0.0000  
>  Mean   : 731595      Mean   :0.4844   Mean   :0.4844  
>  3rd Qu.: 996650      3rd Qu.:1.0000   3rd Qu.:1.0000  
>  Max.   :3604391      Max.   :1.0000   Max.   :1.0000  
> >
>  
> and trying to do a glm like this:
>  
>  
> > logistic.model = glm(formula = as.factor(nextDay) ~ .,
> family=binomial, data=spi[1:50,])
> > summary(logistic.model)
>  
> Call:
> glm(formula = as.factor(nextDay) ~ ., family = binomial, data =
> spi[1:50, 
>     ])
>  
> Deviance Residuals: 
>     Min       1Q   Median       3Q      Max  
> -1.9994  -0.8575  -0.0330   0.7961   1.8376  
>  
> Coefficients: (2 not defined because of singularities)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The answer is most likely here:  You have multicollinearity.

Andy



>                        Estimate Std. Error z value Pr(>|z|)  
> (Intercept)          -3.016e+01  2.934e+01  -1.028   0.3040  
> open                 -2.514e-02  8.939e-02  -0.281   0.7785  
> high                  7.598e-02  1.295e-01   0.587   0.5575  
> low                  -1.065e-01  1.176e-01  -0.905   0.3656  
> close                 5.943e-02  7.995e-02   0.743   0.4573  
> volume                1.960e-04  1.977e-04   0.991   0.3215  
> openInterest          5.728e-05  5.266e-05   1.088   0.2767  
> direction                    NA         NA      NA       NA  
> volatility                   NA         NA      NA       NA  
> volumeXdirection      3.605e-06  3.765e-06   0.958   0.3383  
> volatilityXdirection -5.815e-06  5.708e-06  -1.019   0.3082  
> upDown               -2.561e+00  1.259e+00  -2.034   0.0419 *
> ---
> Signif. codes:  0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1 
>  
> (Dispersion parameter for binomial family taken to be 1)
>  
>     Null deviance: 69.235  on 49  degrees of freedom
> Residual deviance: 55.000  on 40  degrees of freedom
> AIC: 75
>  
> Number of Fisher Scoring iterations: 7
>  
> >
>  
> I am getting NA for direction and volatility.  I got it both 
> before and
> after casting them as.numeric.
>  
> Can anyone tell me what I am doing wrong?
>  
> Thanks
>  
> Stephen
> 
> -- 
> 
> 
> 
> 20/02/2006
>  
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
> 
>