Random Forest Learner Input

[ersan]

I’m trying to run a simple RF to classify a dataset. This is my workflow

Any ideas? My dataset have target value (integer)

[SimonS]

Hi @ersan,

The Random Forest Learner node trains a classification model, i.e., it expects a categorical target column (usually a String). Does your target column represent classes encoded with integers? Then you might use the Number to String node to convert it to a String column. Or do you actually want to train a regression model? Then you need to use the Random Forest Leaner (Regression) node.

I hope this helps,
Simon

[ersan]

Hi Simons,

yep - i have the labels which were represented as integer (now i converted them to string) and trying to predict the label, given features (vector)

I am python based engineer so i’m trying to figure out knime editor, i’m newbie in this context

thanks for the help,

[SimonS]

An example of how to build a simple classification workflow can be found here: https://hub.knime.com/knime/workflows/Examples/04_Analytics/04_Classification_and_Predictive_Modelling/07_Decision_Tree*wjBRbBRnIy6fu4gw

If you want to use random forests, you need to adapt the workflow by replacing the Decision Tree Learner and Predictor nodes with the corresponding Random Forest nodes.

Cheers,
Simon