https://github.com/brendanbignell/ml_kaggle_titanic

Test using LightGBM and FastTree models with GPU acceleration in C#/.NET via ML.NET.
https://github.com/brendanbignell/ml_kaggle_titanic

lightgbm-regressor ml titanic-kaggle

Last synced: about 2 months ago
JSON representation

Test using LightGBM and FastTree models with GPU acceleration in C#/.NET via ML.NET.

• Host: GitHub
• URL: https://github.com/brendanbignell/ml_kaggle_titanic
• Owner: brendanbignell
• Created: 2025-02-05T12:20:15.000Z (4 months ago)
• Default Branch: master
• Last Pushed: 2025-02-05T15:43:02.000Z (4 months ago)
• Last Synced: 2025-02-13T07:37:26.518Z (4 months ago)
• Topics: lightgbm-regressor, ml, titanic-kaggle
• Language: C#
• Homepage:
• Size: 515 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
![image](https://github.com/user-attachments/assets/88fb2c10-30de-4ae4-ad45-d24e8abb5c97)

# ML_Kaggle_Titanic

## Why

The main reason for doing this is to compare building an ML model in C#/.NET vs Python.

## Overview

This is a project to predict the survival of passengers on the Titanic. 

The dataset is from Kaggle and the goal is to predict whether a passenger survived or not. 

The dataset contains 891 rows and 12 columns.

Comparison of LightGBM and FastTree and a deep learning ensemble model consisting of:

	SDCA Logistic Regression

	FastForest

	LBFGS Logistic Regression

Usage of Nvidia GPU for training also tested although on this trivial ammount of data it is not needed.

## Example Output

	Model Comparison Results:  

	Model Type           Accuracy        AUC             F1 Score        Training Time   

	Standard LightGBM    0.8799          0.9585          0.8366          0.20s

	FastTree             0.8878          0.9604          0.8471          0.21s  

	Deep Learning        0.9910          0.9999          0.9883          0.88s