Skip to content
Snippets Groups Projects
Commit 5f934e24 authored by Kyeongpil Kang's avatar Kyeongpil Kang Committed by Joel Nothman
Browse files

[MRG+1] DOC fix sign in GBRT mathematical formulation (#9885)

parent 71841a91
No related branches found
No related tags found
No related merge requests found
Show whitespace changes
Inline Side-by-side
doc/modules/ensemble.rst
+ 2
− 2
View file @ 5f934e24
...@@ -598,7 +598,7 @@ minimize the loss function :math:`L` given the current model ...@@ -598,7 +598,7 @@ minimize the loss function :math:`L` given the current model
.. math:: .. math::
F_m(x) = F_{m-1}(x) + \arg\min_{h} \sum_{i=1}^{n} L(y_i, F_m(x) = F_{m-1}(x) + \arg\min_{h} \sum_{i=1}^{n} L(y_i,
F_{m-1}(x_i) - h(x)) F_{m-1}(x_i) + h(x))
The initial model :math:`F_{0}` is problem specific, for least-squares The initial model :math:`F_{0}` is problem specific, for least-squares
regression one usually chooses the mean of the target values. regression one usually chooses the mean of the target values.
...@@ -614,7 +614,7 @@ loss function: ...@@ -614,7 +614,7 @@ loss function:
.. math:: .. math::
F_m(x) = F_{m-1}(x) + \gamma_m \sum_{i=1}^{n} \nabla_F L(y_i, F_m(x) = F_{m-1}(x) - \gamma_m \sum_{i=1}^{n} \nabla_F L(y_i,
F_{m-1}(x_i)) F_{m-1}(x_i))
Where the step length :math:`\gamma_m` is chosen using line search: Where the step length :math:`\gamma_m` is chosen using line search:
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment