random forest - R randomForest to PMML class index is wrong -


i'm exporting r randomforest model pmml. resulting pmml has class first element of datadictionary element, not true.

is there way fix or @ least increment pmml custom extension elements? way put class index there.

i've looked in pmml package documentation, in pmmltransformations packages, couldn't find there me solve issue.

by pmml class assume mean model type (classification vs regression) in pmml model attributes? if so, not true model type determined data type of first element of datadictionary....these independent. model type determined model type r thinks is. r random forest object determines type thinks (model$type) , model type exported pmml function. if want model type, make sure let r know that...for example, if using iris data set, if predicted variable sepal.length, r correctly assume regression model. if insist on treating classification model, try using as.factor(sepal.length) instead.


Comments

Popular posts from this blog

c# - How to get the current UAC mode -

postgresql - Lazarus + Postgres: incomplete startup packet -

angularjs - ng-repeat duplicating items after page reload -