A Data Science Central Community
Iam currently using R'package to do RF regression method. Dependent variable (Y=1,2,3) and 40 independent variables each sclae from 1 to 5). I have code got node purity for all variables and ordered the variables of importance. As a different approach, I have changed the dependent variable Y= A, B and c instead of 1,2 and 3 with same 40 inde. variables. But in result I got classification and its error rate for A,B and c. additionallty for each variable i got score 9each A,B and C) and Mean Decrease accuracy and Mean decrease Gini.How to intrepret and undertsand the logic.and I got OOB error est.of 33.76%. Could you pl help in this regard.