A collection of small machine learning experiments exploring how different modeling choices, evaluation methods, and assumptions affect results.
Rather than focusing only on model performance, each project tests one specific question or alternative approach.
| No. | Project | Focus |
|---|---|---|
| 01 | Room Occupancy: Time-Aware Evaluation | Random Split vs. Chronological Split |
| 02 | German Credit: Cost-Aware Classification | Default Threshold vs. Cost-Aware Threshold |
| 03 | Olist Customer Satisfaction: Delivery Feature Modeling | Delivery Time vs. Delivery Expectation |
More experiments will be added over time.
Python · pandas · scikit-learn · matplotlib · seaborn · Google Colab