🌾 Data Talk: This one is a tough one...

Multimedia university of kenya

This one is a tough one for sure

Platform · 31 Oct 2022, 17:33 · 18

One of the toughest Zindi competitons this year. Goodluck to the winners. Can't wait to see how you guys solved this problem.

Discussion 18 answers

31 Oct 2022, 17:36

Local CV <-> Public LB <-> Private LB

Upvotes 4

31 Oct 2022, 17:49replied to ASSAZZIN

Dirty data maybe?

Upvotes 0

Muhamed_Tuo

31 Oct 2022, 17:52replied to ASSAZZIN

So so right 🤣

Upvotes 1

31 Oct 2022, 18:01replied to Koleshjr

Test and Train have diffrent distributions .
Small data ( model can't learn easily ), in last year competition (train has 87K rows ) I was able to develop two neural networks that are fairly similar to gbt models. This year's performance is abysmal.
Also Time is important as mentioned here :

Looking at these images, it was evident that the crop got harvested sometime in March or in April starting, which is expected as the wheat crop duration is 100 days, and in most of India, it gets harvested after March. So all the images after harvesting basically capture empty fields, which may be why the composite is not showing vegetation. This means we're basically classifying empty fields, as we lost the important information about the crop by making a composite with the empty fields.

Upvotes 1

31 Oct 2022, 18:06replied to ASSAZZIN

great insights. Also @curiousmonkey7 That was an amazing discussion btw

Upvotes 1

31 Oct 2022, 18:23replied to Koleshjr

Thanks.

Upvotes 0

31 Oct 2022, 18:26replied to ASSAZZIN

"Time is important", @ASSAZZIN ,have you taken time into account while training? because I think it's the same for all images.

Upvotes 1

31 Oct 2022, 19:31replied to curiousmonkey7

No, if I'm not mistaken, All observations has the same date!

Upvotes 0

31 Oct 2022, 19:39replied to ASSAZZIN

Yeah.

Upvotes 0

31 Oct 2022, 17:41

Same here. Nothing seems to be working. I think the dataset could have been better.

Upvotes 0

31 Oct 2022, 17:48replied to curiousmonkey7

The pain of seeing your submissions getting worse and worse when you use what's supposed to be the right cv for such problems 😂😂😂

Upvotes 3

31 Oct 2022, 17:43

how people usually improve their logloss score :

1.50 -> 1.49 ->1.48 ->>>>1.46

People In this Competition :

1.72 ->1.61 ->1.51 ->1.31 ->1.22 ->>1.14

Upvotes 3

31 Oct 2022, 17:45replied to ASSAZZIN

😂😂😂

Upvotes 0