This is my graduation project from Udacity Machine Learning nano degree. It's a multi-label classification project; the goal is to classify outcomes of dogs and cats in animal shelter based on several features such as breeds, colors, intake conditions, ages ... Ideas for this project is from Kaggle Animal Shelter competition: https://www.kaggle.com/c/shelter-animal-outcomes
This project showcases skills I learned from Udacity, including:
- Recognizing data leak problem and provide solutions
- Feature engineering, Exploratory data analysis (EDA) and Preprocessing
- Choosing suitable supervised models based on EDA (Random Forest, Extra Tree Classifier, XGBoost)
- Hyper parameter tuning with grid search + cross validation
- Model evaluation and comparison
- Basic ensemble stacking implementation.
- Improvements
Any feedbacks are welcome!
Quan Tran