| |
 |
 |
|
Last update: February 19, 2014
|
|
|
 |
 |
|
|
Logistic Threshold
For classification models, GeneXproTools allows you to choose different kinds
of rounding thresholds, including fixed and adaptive thresholds.
Logistic regression models, however, use only one type of threshold - the Logistic Threshold,
which is an adaptive
parameter of the
logistic regression model.
The Logistic Threshold is also used in classification models and
is evaluated using the same
logistic regression algorithm that
is used for logistic regression. In both cases, the
number of bins is an adjustable parameter.
When logistic regression runs are converted to classification, the logistic threshold is set
by default, thus keeping the original model thresholds for all the models in the run History.
Note, however, that the only way to keep the original thresholds when
converting classification runs to logistic regression, is when you also
create the classification models with the logistic threshold, as GeneXproTools
updates all the rounding thresholds when converting classification runs to logistic regression.
See Also:
Related Tutorials:
Related Videos:
|
|
|
|
|
