This is the folder for PetFinder.my Adoption Prediction competition, team "BaMoOaAl"
Team Components:
- https://www.kaggle.com/alessandrosolbiati, SolbiatiAlessandro
- https://www.kaggle.com/oanaflorescu, flores-o
We will follow along Standford CS231-n assignments and implement them here on the competition.
ROADMAP:
Models:
- KNN: best validation score: 0.20 + 0.14 public LB score: 0.279 BIMODEL
- NB: best validation score: 0.10249 public LB score: 0.172
- implement SVM: best validation score: public LB score:
- implement NN (ResNet transfer learning)
- LGBM: best validation score: 0.17435 public LB score: 0.278
- CATBOOST: best validation score: 0.20133 public LB score: 0.349
Framework:
- write standard PredictiveModel
- write test
- write benchmark/execution scripts ( we are using notebooks )
- write docs with model performance and insight
- add code coverage
Exploratory Data Analysis + Feature Engineering
- Adoption Speed
- Name
- Age
- Breed
- Color
- Size
- Country
- Images