// the find
wepe/Kaggle-Solution
Solution Code for Kaggle Competition
A personal collection of Kaggle competition solutions from roughly 2014-2016, covering classic ML tasks like digit recognition, diabetic retinopathy detection, and product classification. It's a historical snapshot of one competitor's work, not a framework or reusable library.
Each competition lives in its own directory with a readme explaining the approach, which makes it easy to find the relevant code. The diabetic retinopathy section includes a full CNN pipeline with separate preprocessing, augmentation, and prediction scripts — reasonable structure for its era. The Otto Group section shows ensemble thinking by keeping each classifier (RF, GBM, ExtraTrees, AdaBoost) in its own file, making it easy to see what was tried.
Last updated in 2021 but the code itself is from ~2014-2016 — Keras/TF versions, sklearn APIs, and data pipeline patterns are all stale. There are no requirements files, no environment specs, and no data download instructions, so actually running any of this requires archaeology. The README is four words. This is a personal archive, not a reference implementation — anyone expecting documented, reproducible experiments will be disappointed.