M1 Machine Learning Project (Graded)

This repository contains clean, publishable material for multiple ML notebooks with their required datasets.

Project Objectives

1. Graded project notebook: predict genre from movie/series metadata and text features.
2. TD Project 01 notebook: introductory data manipulation and analysis tasks.
3. TD Project 02 notebook: classification workflow on breast cancer data.
4. TD Project 03 notebook: predictive maintenance classification workflow.
5. TD Project 04 notebook: run classification experiments on heart disease, stars, and glass datasets.

Repository Structure

• TD_2026_S2_M1_ML_Project_Graded.ipynb — complete notebook (problem framing, preprocessing, modeling, evaluation)
• imdb_descr_titles.csv — input dataset
• TD_2026_S2_M1_ML_Project_04.ipynb — TD notebook
• heart_disease_classification.csv — dataset used in TD Project 04
• stars_nasa_classification.csv — dataset used in TD Project 04
• glass_classification.csv — dataset used in TD Project 04
• TD_2026_S2_M1_ML_Project_01.ipynb — TD notebook
• EU_countries.csv — dataset used in TD Project 01
• TD_2026_S2_M1_ML_Project_02.ipynb — TD notebook
• breast_cancer.csv — dataset used in TD Project 02
• TD_2026_S2_M1_ML_Project_03.ipynb — TD notebook
• predictive_maintenance.csv — dataset used in TD Project 03

Methodology (Notebook Summary)

The notebook follows a full ML workflow:

1. Problem framing (multiclass supervised classification)
2. Data preprocessing
   • target construction from genres
   • text feature engineering (title + description)
   • categorical preprocessing (One-Hot Encoding)
   • numeric imputation
3. Modeling
   • DummyClassifier (baseline)
   • LogisticRegression
   • LinearSVC
   • MultinomialNB (text baseline)
4. Evaluation
   • cross-validation metrics
   • test metrics
   • comparison table and short commentary

Recommended Environment

• Python 3.10+
• numpy
• pandas
• scikit-learn
• jupyter

Install dependencies (if needed):

pip install numpy pandas scikit-learn jupyter

How to Run

A) Graded notebook

1. Keep imdb_descr_titles.csv in the same directory as the notebook.
2. Open TD_2026_S2_M1_ML_Project_Graded.ipynb.
3. Run all cells sequentially from top to bottom.

B) TD notebook (Project 04)

1. Keep these files in the same directory:
   • heart_disease_classification.csv
   • stars_nasa_classification.csv
   • glass_classification.csv
2. Open TD_2026_S2_M1_ML_Project_04.ipynb.
3. Run all cells sequentially from top to bottom.

C) TD notebook (Project 01)

1. Keep EU_countries.csv in the same directory as the notebook.
2. Open TD_2026_S2_M1_ML_Project_01.ipynb.
3. Run all cells sequentially from top to bottom.

D) TD notebook (Project 02)

1. Keep breast_cancer.csv in the same directory as the notebook.
2. Open TD_2026_S2_M1_ML_Project_02.ipynb.
3. Run all cells sequentially from top to bottom.

E) TD notebook (Project 03)

1. Keep predictive_maintenance.csv in the same directory as the notebook.
2. Open TD_2026_S2_M1_ML_Project_03.ipynb.
3. Run all cells sequentially from top to bottom.

Reproducibility

The notebook uses fixed random seeds in key steps (e.g., stratified split / CV settings where applicable) to keep results stable across runs.

Notes

• The dataset size is compatible with standard GitHub limits (well below 100 MB).
• The repository is intentionally minimal and focused on the published notebooks + their datasets.