BINARY CLASS PROBLEM
USING DECISIONTREEREGRESSOR FUNCTION
Target variable is dichotomous (yes or No-1 or 0)
CRITERION – MEAN SQUARED_ERROR(MSE)
Variance Reduction
Result
CRITERION: MEAN ABSOLUTE ERROR (MAE)
Find MAE for each node
CRITERION: FRIEDMAN MSE
CRITERION: POISSON
MULTI-CLASS PROBLEMS
Criterion: Mean Squared_Error (MSE)
We use iris data
criterion – absolute_error (mae)
Criterion – friedman_mse and value
criterion- poisson and mean value
Summary
