-
🔭 I’m currently working on [ R and Python]
-
🌱 I’m currently working on Feature & Target Engineering. Dimension reduction methods focus on reducing the feature space: "live with your data before you plunge into modeling"
Although not always a requirement, transforming the response variable can lead to predictive improvement. For Example, ordinary linear regression models assume that the prediction errors are normally distributed. This is normally fine, but when the prediction target has heavy tails (outliers) or is skewed in one direction. In these cases, the normality assumption likely does not hold.
We update the repo soon possible :
In 2023, there will be three times as many linked gadgets as there are people on the planet. By 2023, there will be 1.6 networked mobile devices and connections per person.**
|
