From: Building efficient fuzzy regression trees for large scale and high dimensional problems
Parameter | MLlib DRT | DFRT |
---|---|---|
Fuzzy discretization | ||
\(\lambda _{\text{disc}}\) | – | 1% |
\(\epsilon _{\text{disc}}\) | – | 0.0001 |
\(\gamma _{\text{disc}}\) | – | 0.1% |
Regression tree | ||
---|---|---|
Impurity | Variance | – |
\(\beta\) | 15 | 15 |
L (maximum number of bins) | 100 | 100 |
\(\lambda _{\text{learn}}\) | 0.5% | 0.5% |
\(\epsilon _{\text{learn}}\) | 0.0001 | 0.0001 |