search

AlineTalhouk / splendid

Supervised Learning Ensemble for Diagnostic Identification
https://alinetalhouk.github.io/splendid/
Other
1 stars 0 forks source link

Calibration and updating #15

Closed AlineTalhouk closed 5 years ago

AlineTalhouk commented 7 years ago

with R code and very well explained. Review chapter 6 in particular. Very interesting read and well explained ftp://ftp.esat.kuleuven.be/pub/SISTA/ida/reports/14-224.pdf

AlineTalhouk commented 7 years ago

I will put some of these ideas to practice tomorrow. I just wanted to share the material

Dustin21 commented 7 years ago

@dchiu911: Just wondering if there was a change made to the predict() function? After obtaining the top classifiers, Aline and I have tried predict(), which returns the following error:

Error in eval(expr, envir, enclos) : object 'HLA.DRA' not found.

I've tried prediction(), which works fine, but we're just wondering if a change was made here recently?

dchiu911 commented 7 years ago

I'm not sure what code you ran to produce the error, but I believe it relates to using the maboost method. Notice in prediction.maboost I need to apply make.names on all the variables of data, to make sure that names such as "HLA-DRA" with the restricted string "-" is turned into a period. So if you're trying to run a predict() outside of splendid, try make.names