Titanic Survival Prediction Model Information

How to Use This Page

This page gives you a peek under the hood of the model I used for this project. It provides a detailed overview of the XGBoost model powering the Titanic survival predictions. You'll find the model's configuration, the features it considers, and how it made its predictions.

Executive Summary

Below is a 100-foot view of the model and the dataset.

Model

XGBoost

Features

Version

v1.0

Trees Built

100

Tree Depth

Learning Rate

0.1

What This Model Does

This machine learning model predicts whether a passenger on the Titanic survived based on their personal characteristics and travel details. It's a binary classifier, meaning it makes yes/no predictions. In the case of the Titanic dataset, it predicted if passengers survived.

Features the Model Considers

The model analyzes 13 different passenger characteristics to make predictions:

Personal Information

Age: Passenger's age and age group
Sex: Gender of the passenger
Title: Social title (Mr., Mrs., Dr, etc.)

Ticket Details

Passenger Class: 1st, 2nd, or 3rd class
Fare: Ticket price and price category
Cabin: Whether passenger had a cabin
Embarked: Port embarked from

Family Information

Family Size: Total family members aboard
Siblings/Spouse: Number aboard
Parents/Children: Number aboard
Is Alone: Traveling solo or not

How the Model Works

The Extreme Gradient Boosting (XGBoost) model works like a committee of decision-makers to return optimal predictions, following a set of prescribed steps:

It creates 100 decision trees to tackle the prediction task.
Each tree can ask up to 3 questions about a passenger.
The model learns at a rate of 0.1. A slower learning rate improves accuracy.
All trees 'vote' on the final prediction, with later trees correcting earlier trees' mistakes.

Download Output Data

If you need to use the input and output data to build a dashboard using a different tool, you can download the data.

Export Data

Super nerdy details 🤓

Feature Names

[
  "Pclass",
  "Sex_Encoded",
  "Age",
  "SibSp",
  "Parch",
  "Fare",
  "Embarked_Encoded",
  "FamilySize",
  "IsAlone",
  "Title_Encoded",
  "HasCabin",
  "FareBin_Encoded",
  "AgeGroup_Encoded"
]

Model Parameters

{
  "objective": "binary:logistic",
  "base_score": null,
  "booster": null,
  "callbacks": null,
  "colsample_bylevel": null,
  "colsample_bynode": null,
  "colsample_bytree": 1.0,
  "device": null,
  "early_stopping_rounds": null,
  "enable_categorical": true,
  "eval_metric": "logloss",
  "feature_types": null,
  "feature_weights": null,
  "gamma": null,
  "grow_policy": null,
  "importance_type": null,
  "interaction_constraints": null,
  "learning_rate": 0.1,
  "max_bin": null,
  "max_cat_threshold": null,
  "max_cat_to_onehot": null,
  "max_delta_step": null,
  "max_depth": 3,
  "max_leaves": null,
  "min_child_weight": null,
  "missing": NaN,
  "monotone_constraints": null,
  "multi_strategy": null,
  "n_estimators": 100,
  "n_jobs": -1,
  "num_parallel_tree": null,
  "random_state": 42,
  "reg_alpha": null,
  "reg_lambda": null,
  "sampling_method": null,
  "scale_pos_weight": 1.0,
  "subsample": 1.0,
  "tree_method": null,
  "validate_parameters": null,
  "verbosity": null
}

XGBoost Classifier Model Info