Kaggle Competition Silver Medal: Child Mind Institute-Problematic Internet Use

I am excited to announce my Silver Medal achievement in the Child Mind Institute Problematic Internet Use Kaggle competition, which ran from October to December 2024.

The project’s goal was to develop a model predicting problematic internet usage patterns in children and adolescents by analyzing physical activity data. This initiative aims to identify early signs of problematic internet use and promote healthy digital habits.

Technical Implementation:

• Tools: Python, LightGBM, XGBoost, CatBoost, Polars, KNNImputer
• Dataset: Healthy Brain Network (HBN) dataset containing accelerometer and health-related data
• Solution:
  1. Efficient data preprocessing using Polars, with KNNImputer for missing values
  2. Ensemble learning combining LightGBM, XGBoost, and CatBoost models
  3. StratifiedKFold cross-validation for balanced prediction
  4. Solution notebook

Results:

• Public Score: 0.460, Private Score: 0.454
• Achieved Silver Medal, ranking in Top 3%