https://github.com/thiraput01/cibmtr-equity-in-post-hct-survival-predictions
Notebooks for Kaggle competition
https://github.com/thiraput01/cibmtr-equity-in-post-hct-survival-predictions
catboost-regressor custom-metrics data-cleaning data-processing data-science k-fold-cross-validation lightgbm-regressor machine-learning optuna regression-models xgboost-regressor
Last synced: 2 months ago
JSON representation
Notebooks for Kaggle competition
- Host: GitHub
- URL: https://github.com/thiraput01/cibmtr-equity-in-post-hct-survival-predictions
- Owner: Thiraput01
- Created: 2024-12-23T14:01:24.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2025-01-25T13:34:43.000Z (5 months ago)
- Last Synced: 2025-03-23T23:14:33.508Z (3 months ago)
- Topics: catboost-regressor, custom-metrics, data-cleaning, data-processing, data-science, k-fold-cross-validation, lightgbm-regressor, machine-learning, optuna, regression-models, xgboost-regressor
- Language: Jupyter Notebook
- Homepage:
- Size: 508 KB
- Stars: 5
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Equity in Post-HCT Survival Predictions
This repository contains a project that focuses on predicting post-hematopoietic cell transplantation (HCT) survival outcomes while considering equity across patient groups.
This notebook is used for [CIBMTR - Equity in post-HCT Survival Predictions](https://www.kaggle.com/competitions/equity-post-HCT-survival-predictions/overview) Kaggle competition
## Project Overview
The project explores:
- Predictive modeling of post-HCT survival.
- Incorporating fairness into survival predictions.
It includes the following key components:
1. **Data Preprocessing:** Handling and preparing the dataset for analysis.
2. **Modeling:** Developing predictive models to estimate survival outcomes.
## Key Features
- End-to-end workflow for survival prediction modeling.
- Tools to assess model performance and fairness metrics.
- Evaluation Metrics:
- Applied the Kaplan-Meier method for survival analysis.
- Used Concordance Index (C-index) to evaluate predictive performance.
