https://github.com/programmer-ke/icr-age-related-kaggle-competition

ICR - Identifying Age-Related Conditions
https://github.com/programmer-ke/icr-age-related-kaggle-competition

Last synced: 9 days ago
JSON representation

ICR - Identifying Age-Related Conditions

• Host: GitHub
• URL: https://github.com/programmer-ke/icr-age-related-kaggle-competition
• Owner: programmer-ke
• Created: 2023-08-11T07:12:22.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-10-26T09:16:15.000Z (about 1 year ago)
• Last Synced: 2023-10-26T10:29:46.402Z (about 1 year ago)
• Language: Jupyter Notebook
• Size: 1.35 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Kaggle Competition [ICR Identifying Age-Related Conditions][1]

This repo has the notebooks used for the competition.

The best performing notebook on the private leaderboard was 

[this](notebooks/kfold-ensemble-gb-rf-nn.ipynb), scoring 0.28 in the

public leaderboard and 0.44 in the private leaderboard, almost getting

a bronze level position.

The ones I submitted however did not perform as well. Those were

[this](notebooks/hgb-kfold.ipynb) (0.26 public, 0.58 private) and

[this](notebooks/ensemble-them-all.ipynb) (0.6 public, 0.91 private).

Overall, it was quite an interesting competition. A small dataset

(which I understand is common with health records..),

how easy it was to overfit, how much even a [small shift][2] on  a single

data point's predictions impacted the overall metric, etc..

The winner's solution is explained [here][3].

I'll try document any learnings that will emerge.

[1]: https://www.kaggle.com/competitions/icr-identify-age-related-conditions

[2]: https://www.kaggle.com/competitions/icr-identify-age-related-conditions/discussion/430475

[3]: https://www.kaggle.com/competitions/icr-identify-age-related-conditions/discussion/430843