Update the user guide #199
Conversation
|
|
||
| # Create the foreshadow object | ||
| shadow = fs.Foreshadow(X_preparer=x_processor, y_preparer=y_processor, estimator=XGBRegressor()) | ||
| Foreshadow uses a machine learning model to power the auto intent resolving step. As a user, you may not agree with the decision made by Foreshadow. The following APIs allow you to inspect the decisions and change them if you have a different opinion. |
There was a problem hiding this comment.
We might want to add a section to briefly introduce the major functionality of Foreshadow at the beginning and point that one of the features is automatic intent resolving (we might need to explain what we mean by intent as well). With that foundation established, it will be easier to just say the decisions can be changed by users here.
doc/users.rst
Outdated
|
|
||
| First import foreshadow | ||
| In this tutorial, we will go through a typical ML workflow with Foreshadow using a subset of the `adult data set <https://archive.ics.uci.edu/ml/datasets/Adult>`_ from the UCI machine learning repository. |
There was a problem hiding this comment.
Minor quibble: Wonder if more common datasets out there (like Boston or IRIS used by most tutorials) that people might be more familiar with are a better fit here. Sklearn datasets are probably a good place to dig.
Shouldn't be blocking though - adult dataset is a good representation of what you might see in real life.
There was a problem hiding this comment.
I actually spotted a bug using the Iris data set with intent override. The fix is already published to a PR. Thanks for pointing this out!
jzhang-gp
force-pushed
the
update_docs
branch
from
f393c90 to
b512875
Compare
Description
Update the user doc.