Skip to content

get the calibration model and add a predict() function #10

Description

@mdedonno1337

Hi,

I was wandering if you'd like to have way to compute the calibration model (the parameters), and be able to use a 'predict()' function to predict, based upon the calibration done on a known dataset, the calibrated values for new values (that are not in the calibration set).

It's not very difficult with the current state to compute the calibrated values for new input LRs, but you have to understand how the calibration process is done and code yourself the calibration parameters and the apply-calibration functions.

I can do this for the pava (calibration.set() function) (and the logistic calibration if I add it later on).

Marco

EDIT:
The calibrated values are, in some circonsances a bit difficult to direclty relate to the input data, in particular if the input data is not sorted. The predict() function could help I think.

> lrss
 [1] 316227.76602    562.34133    177.82794     56.23413  56234.13252
 [6] 100000.00000   1778.27941  31622.77660  17782.79410  10000.00000
> lrds
 [1] 1000.000000   56.234133  100.000000 3162.277660    5.623413   10.000000
 [7]  316.227766   31.622777    3.162278   17.782794

> calibrate.set( lrss, lrds )
$LR.cal.ss
 [1]   1   1   1   1 Inf Inf Inf Inf Inf Inf

$LR.cal.ds
 [1] 0 0 0 0 0 0 1 1 1 1

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions