ERROR: Unexpected Column: '' (Line 1, Column 1)

How we should treat null values in a category parameter?

Since we have null values in Age we need to define an strategy to treat them:

1. We can fill the NaN with the mean or the median
2. We can fill the NaN with random numbers conserving the mean and the median.
3. We can delete the rows.

We find that other competitors with apparently worse data preprocessing (label encoding for preprocesing the database, not normalizing numeric variables, etc...) have quite better performance for their models. Things that maybe explain this:

• Null values in the categoric parameters must be fill with the mode of those variables.
• Not join the luxury expenses

We have to process the parameter in the ritgh way, this means not just filling the null values with zeros but like we stablish in the notebook:

• Zeros for the cryosleepers
• And the mean for each group fo VIP and no VIP passengers.

Do the preprocessing to the test data.