How to create a target from multiple columns to train a classification model?

Question

DataFrame head sample: Where I need to create a new feature of the name category which will contain sarcastic category data like data from the last six columns should combine to make the new column and the values in that column should be labeled as per the column's name of the last five columns.

Answer 1

Use df.loc['category A' = 1, 'category'] = 'category A' for all possible categories:

categories = ['sarcasm', 'irony', 'satire', 'understatement', 'overstatement', 'rhetorical_question']
for category in categories:
    df.loc[df[category] == 1, 'category'] = category