iX Mobile Banking Prediction Challenge
Helping Africa
$3 495 USD
Completed (over 5 years ago)
Prediction
349 joined
103 active
Start
17 May 21
Close
13 Jun 21
Reveal
13 Jun 21
Should we use country_code and region as categorical features or numerical ones?
Data · 9 Jun 2021, 03:54 · 2
In reality, country_code and region are categorical features. But in our dataset, they are encoded as numerical ones. So, in which encoding format should we input these features for helping the model to better generalize?
They are categorical no matter the format they come in because each country has a unique number.
We all understand that they are categorical features. But passing them to the model as numerical features, the model will try to compute weight like someone belonging to a country with a higher country_code/region will have more/less chances to use mobile/internet banking. Anyway, thanks for sharing ...