Some weeks in the past, I printed a publish on LinkedIn.
The publish was primarily based on the next determine, evaluating the predictions made by two fashions: Linear Regression, and CatBoost.
The gist of the publish was {that a} gradient-boosting mannequin like CatBoost appears to offer a extra “cheap” interpretation of the connection between a predictor and the goal variable (particularly, home situation and home value).
Certainly, many relationships in nature are non-linear.
The publish obtained some objections, amongst which the next remark stood out for the big variety of likes:
This began a dialogue, and I discovered the next remark (written by the identical writer…
