Skip to content

erikycd/Titanic_survivor_prediction

Repository files navigation

Titanic_survivor_prediction

A Machine Learning-based guide to predict the survival rate of the Titanic disaster

From the classic Titanic machine learning (ML) competition on Kaggle website. The very first steps to dive into ML competitions and familiarize with prediction systems.

The approach shown here makes use of classic ML-based techniques to create a model that predicts which passengers survived the Titanic shipwreck. Python file.

Basically, the process involves:

* Feature engineering
  * Data analysis
  * Age prediction
  * Fares estimation
* Data cleaning
* Data normalization and scaling
* Automatic classification

Training data description.

Composed of 891 rows with individual information each that includes: PassengerID, Survived value, Pclass, Name, Sex, Age, SibSp, Parch, Ticket, Fare, Cabin and Embarked.

train

Test data description

Composed of 418 rows with individual information each that includes: PassengerID, Pclass, Name, Sex, Age, SibSp, Parch, Ticket, Fare, Cabin and Embarked.

test

Unbalanced data sorted by Sex vs Survived

unbalanced

Feature correlation map

The feature correlation map shows if there exists any statistical association between two variables. The closer value to 1 or -1, the higher the correlation. For instance, th e Survived feature is highly correlated to Pclass and Fares

correlation

Feature correlation map after preprocessing

Now, The Survived feature is highly correlated to Pclass and Sex

correlation

Classification results with sklearn library

Different classification methods were applied in the training set.

classification

About

A ML guide to predict the survival rate in the Titanic disaster

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages