my predicting churn value is 0 or 1

sample submission file is also contain only 0 and 1

Check 'evaluation'. Submission should look like this:

user_id CHURN 00001dbe00e56fc4b1c1b65dda63de2a5ece55f9 0.98 000055d41c8a62052dd426592e8a4a3342bf565d 0.12 000081dd3245e6869a4a9c574c7050e7bb84c2c8 0.37

From the "Evaluation" page:

"The values can be between 0 and 1, inclusive. Where 1 indicates the customer churned and 0 indicates the customer stayed with Expresso.

The submission file is large, please expect a longer scoring time.

Your submission should look like:

user_id CHURN 00001dbe00e56fc4b1c1b65dda63de2a5ece55f9 0.98 000055d41c8a62052dd426592e8a4a3342bf565d 0.12 000081dd3245e6869a4a9c574c7050e7bb84c2c8 0.37"

Use something like "model.predict_proba(test_set)[:,1]" as Johannes suggested.

Thank you dude I got 90% accuracy

Big thanks!!!

Do i need to round up values (of Target)?

Not sure - I didn't round anything

No problem. I had the same trouble for my first submissions and it took me some time to figure out not to round them.

ok. Thanks

This one was very helpful to me I had predictions of 0 and 1 which gave me 87% so after the change now I got 91%