Feature As Predictor: Add documentation #3
Conversation
|
|
||
| 2. Enable transformation. When transformation is applied, the model constructed by the widget is equivalent to a logistic or linear regression model. | ||
|
|
||
| This feature is disable when the transformation is not applicable or enforced. |
|
|
||
|  | ||
|
|
||
| How well is the price of a house based on the number of rooms? Here we can (and must) enable transformation because the scale of the feature's values is very different from the scale of the target variable. Note that this measure the fit rather than predictive power because we do not test the model on unseen data. |
There was a problem hiding this comment.
How well is the price of a house based on the number of rooms? --> How well does the price of a house correlate with the number of rooms? (or something similar)
There was a problem hiding this comment.
Note that this measure the fit rather than predictive power --> Note, that this measures the fit rather than the predictive power
|
|
||
|  | ||
|
|
||
| How well can we predict the heart disease based on the maximal heart rate? Here we connect the *Feature as Predictor* widget to the *Test and Score* widget to evaluate the model on unseen data. In this case, Feature as Predictor provides a learning algorithm that fits a logistic regression model to the feature's values. |
There was a problem hiding this comment.
At one point, you'll have to explain the difference between maximum and maximal.
There was a problem hiding this comment.
Maximal is an adjective (maximal speed), and maximum is adjective or noun (maximum speed, global maximum).
Maximum heart rate should thus be OK, but I tend to always use maximal when I need an adjective. Probably because it's like this is Slovenian, where "maksimum" cannot be an adjective.
janezd
force-pushed
the
feature-as-predictor
branch
from
3abe287 to
b1bff04
Compare
No description provided.