Played around with the Titanic Dataset to predict whether a person survives or not. (https://www.kaggle.com/c/titanic)
- Dropped 'Ticket' and 'Cabin' features
- Used an unknown token "U" for null points in 'Embarked' variable
- Used median to fill the null points in 'SibSp', 'Parch', 'Fare' and 'Age'
- LogisticRegression
- Accuracy : 76.31 (Public Leaderboard)