I am so Sorry and i know were are both in-need of the 2000 points, but please can someone help on how to treat the country column to get better scores, I tried removing it and still get worst results, i also tried to find how the test countries correlate with the train countries and still get worse results. Any suggestions are all welcome
hi @BrightXO!
Sorry for the lateness.
You can use either https://github.com/jeongyoonlee/Kaggler or https://github.com/viktorsapozhok/cafeen/blob/master/cafeen/steps.py
I am using FrequencyEncoder from Kaggler and it pretty well improved my results...
Best,
@CapitainData