add quantile regression example #608
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
| @@ -0,0 +1,621 @@ | |||
| { | |||
There was a problem hiding this comment.
| @@ -0,0 +1,621 @@ | |||
| { | |||
There was a problem hiding this comment.
The first paragraph says "Normal" and the second says "normal".
Then I would use "For example, in medical research, pathologies or potential health risks occur at high or low quantiles, for instance, overweight and underweight." (too many commas, but I think it's clearer).
"In some other fields like ecology quantile" -> " In some other fields like ecology, quantile"
Reply via ReviewNB
| @@ -0,0 +1,621 @@ | |||
| { | |||
There was a problem hiding this comment.
"assymetric paramter" -> "assymetry parameter". After that, I would add a colon ":" or a period ".", but not a comma ",".
Also, does the "kappa" have an interpretable meaning? Asking because honestly I don't know
Reply via ReviewNB
There was a problem hiding this comment.
Just saw a typo: "There is at least two alternative parameterizations" -> "There are..."
| @@ -0,0 +1,621 @@ | |||
| { | |||
There was a problem hiding this comment.
I would add a formula showing like this
$$
\kappa = (\frac{q}{1 - q}) ^ 0.5
$$
So it looks like it's the square root of the logit of q? I don't know if it adds any value, I just realized about it
Reply via ReviewNB
| @@ -0,0 +1,621 @@ | |||
| { | |||
There was a problem hiding this comment.
I would add one more sentence saying the assymetriclaplace family allows the model to account for the increased variability in BMI as the age increases, while for the Gaussian family that variability always stays the same.
Reply via ReviewNB
|
Overall it looks really good. I would consider the following points
|
Codecov Report
@@ Coverage Diff @@
## main #608 +/- ##
=======================================
Coverage 82.84% 82.84%
=======================================
Files 38 38
Lines 3026 3026
=======================================
Hits 2507 2507
Misses 519 519 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
Closes #593