Question

I've been using modFit (part of the R package, FME) on a recent project. I am interested in obtaining confidence intervals for the model predicted parameters, but I cannot figure out how to do so. When I look at the summary from the modFit routine, I am given the standard error and t-value for each parameter. I am also given a parameter correlation table.

How can I extract a confidence interval for the parameters with this information? I feel like I'm over-complicating the task, but I am just stuck. Any help would be greatly appreciated.

This is this summary information I have to work with (I hope it doesn't just add more confusion to my question, so disregard if need be):

Parameters:
      Estimate Std. Error t value Pr(>|t|)    
Mb1  1.7022419  0.0022914  742.89   <2e-16 ***
M1b  2.0015269  0.0024875  804.64   <2e-16 ***
Mb2  0.4998066  0.0015215  328.50   <2e-16 ***
M2b  0.5079805  0.0011033  460.43   <2e-16 ***

---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

Residual standard error: 0.07965 on 67 degrees of freedom

Parameter correlation:
           Mb1     M1b     Mb2       M2b    
Mb1   1.000000  0.8377 -0.6335  0.006307 
M1b   0.837665  1.0000 -0.6208 -0.163344 
Mb2  -0.633522 -0.6208  1.0000  0.529634 
M2b   0.006307 -0.1633  0.5296  1.000000 
Was it helpful?

Solution

You can try bootstrap package boot it will build you confidence interval based on re-sampling

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top