Xgboost: predict after cross-validation using xgboost [question]
This is my first trial with xgboost (very fast!). But I'm a little bit confused .
In fact, I trained a model using xgb.cv as follows:
xgbmodel=xgb.cv(params=param, data=trainingdata, nrounds=100, nfold=5,showsd=T,metrics='logloss')
Now I want to predict with my test set but xgbmodel seems to be a logical value (TRUE in this case)
How could I predict after cv? Should I use xgb.train then?
HR
RanaivosonHerimanitra
All 3 comments
Yes, the xgb.cv does not return the model, but the cv history of the process. Since in cv we are training n models to evaluate the result.
A normal use case of cv is to select parameters, so usually you use cv to find a good parameter, and use xgb.train to train the model on the entire dataset
tqchen
on 1 Nov 2014
Ok, It's more clear now
RanaivosonHerimanitra
on 2 Nov 2014
Hi,
There is a parameter prediction=TRUE in xgb.cv, which returns the prediction of cv folds. But it is not clear from the document that for which nround, the predictions are returned?
vikasnitk85
on 26 Jan 2016
Related issues
frankzhangrui
·
3Comments
matthewmav
·
3Comments
trivialfis
·
3Comments
pplonski
·
3Comments
ivannz
·
3Comments
Most helpful comment
Yes, the xgb.cv does not return the model, but the cv history of the process. Since in cv we are training n models to evaluate the result.
A normal use case of cv is to select parameters, so usually you use cv to find a good parameter, and use xgb.train to train the model on the entire dataset