This repository contains my attempts at learning various techniques and methods to solve a ML classification problem. The problem chosen is a competition from Kaggle, the Shelter Animal Outcomes.
The tools used are the excellent scikit learn, jupyter, matplotlib, seaborn and pandas.
The problem is to predict the probabilities of an animal getting adopted, transfered, death from natural causes, euthanasia, or returned to it's owner. The following features are given to us age, breed, color, gender, spayed/neutered/intact, whether it's a cat/dog, has a name, etc.
The suggested reading order would be:
- Visualization
- Data cleaning
- Random forests
- Logistic Regression
- Naive Bayes
- Decision Tree
- SVM
- KNN
- ExtraTreesClassifier
- AdaBoost
- GradientTreeBoosting
- Bagging
You can look at the model evaluations here