Methodology:

Filtered out some columns which seemed less impactful or had too many NaNs.
Used most frequent value simple imputer for filling the rest.
One hot encoding used for non-integer columns.
Fit model using Multinomial Naive Bayes.

FILE TO BE CHECKED IS result.csv