ambitious-octopus / SimAnn

Research unit for the study of the BAM Model
2 stars 0 forks source link

controllare l overfitting del random forest #4

Closed StefSeSi closed 4 years ago

StefSeSi commented 4 years ago

visti alla veloce su internet: n_estimators: The more trees, the less likely the algorithm is to overfit. ... max_features: You should try reducing this number. ... max_depth: This parameter will reduce the complexity of the learned models, lowering over fitting risk. min_samples_leaf: Try setting these values greater than one.

direi pure di provare una cross validation ma sempre c'è il problema che le y variano poco (al 90% stabili intorno a un valore) ... si potrebbe agirare con delle partizioni di crossvalidation "ritoccate": ben variate sulla y (...)

ambitious-octopus commented 4 years ago

Thank you, I will directly implement a Grid_search.