Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[Breaking] Change default evaluation metric for classification to logloss / mlogloss (#6183)

Browse Source 
* Change DefaultEvalMetric of classification from error to logloss

* Change default binary metric in plugin/example/custom_obj.cc

* Set old error metric in python tests

* Set old error metric in R tests

* Fix missed eval metrics and typos in R tests

* Fix setting eval_metric twice in R tests

* Add warning for empty eval_metric for classification

* Fix Dask tests

Co-authored-by: Hyunsu Cho <chohyu01@cs.washington.edu>
This commit is contained in:
Christian Lorentzen
2020-10-02 21:06:47 +02:00
committed by GitHub
parent e0e4f15d0e
commit cf4f019ed6
18 changed files with 56 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
plugin/example/custom_obj.cc
View File

@@ -56,7 +56,7 @@ class MyLogistic : public ObjFunction {
}
}
const char* DefaultEvalMetric() const override {
return "error";
return "logloss";
}
void PredTransform(HostDeviceVector<bst_float> *io_preds) override {
// transform margin value to probability.

2
plugin/updater_oneapi/regression_loss_oneapi.h
View File

@@ -103,7 +103,7 @@ struct LogisticRegressionOneAPI {
// logistic loss for binary classification task
struct LogisticClassificationOneAPI : public LogisticRegressionOneAPI {
static const char* DefaultEvalMetric() { return "error"; }
static const char* DefaultEvalMetric() { return "logloss"; }
static const char* Name() { return "binary:logistic_oneapi"; }
};