GeoAI Agricultural Plastic Cover Mapping with Satellite Imagery by ITU
Helping Italy
1 000 CHF
Challenge completed ~1 year ago
Prediction
365 joined
78 active
Start
Jun 12, 24
Close
Nov 11, 24
Reveal
Nov 11, 24
Cross Validation
Help · 30 Aug 2024, 18:31 · 6
What type of cross validation are you using. This is because my CV score is really far from what I'm having on the Public Leaderboard. I'm probably doing something wrong.
I am using StratifiedKFold
I don't think there is anything wrong with the cross validation you are using. Check if your target labels are 1 and 2 instead of 0 and 1.
@Nyatuka Thank you for your response. I changed the 2s to 0s as defined in the variable definition file "TARGET,"This is what you are predicting, 1 denotes agricultural plasic cover, 2 denotes non-agricultural plastic cover".
I have attached the link to my notebook for you to have a look.
https://drive.google.com/file/d/1_u4Ja_Mcas-793mrz7mwhSqLcwB1Jsj4/view?usp=drive_link
Thank you!
I meant the right labels are 1 and 2, because when using some models for binary classification they requre our labels to be 0 and 1, I just thought you hadn't reverted to the original labels.
I reverted the original labels to 0s and 1s.
Hey could any one please tell me how to make IDs for these testing file i making worng and wrong again