Machine Learning Notebooks: Decision Trees + MNIST

Jupyter notebooks demonstrating Decision Tree classification and MNIST digit recognition. [file:28][file:29]

Contents

Notebook	Description	Dataset
decision-_tree.ipynb	Decision Tree classifier implementation, training, evaluation	Custom/synthetic data
tmnst.ipynb	MNIST handwritten digits (0-9) classification	MNIST dataset (~60k images)

Quick Start

# Clone repo & install dependencies
git clone <your-repo-url>
cd repo-name

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# .venv\Scripts\activate  # Windows

# Install packages
pip install jupyter pandas numpy scikit-learn matplotlib seaborn pillow

# Launch notebooks
jupyter notebook

Tech Stack

pandas | numpy | scikit-learn | matplotlib | seaborn | jupyter | PIL

Key Concepts Covered

Decision Tree Notebook [file:28]

• Tree construction algorithm
• Entropy/Gini impurity
• Hyperparameter tuning (max_depth, min_samples_split)
• Visualization of decision boundaries
• Cross-validation scoring

MNIST Notebook [file:29]

• Image preprocessing (28x28 → flatten/reshape)
• Data splitting (train/test)
• Model training on 60k+ samples
• Confusion matrix & classification report
• Digit prediction visualization

Results Preview

Decision Tree: Accuracy ~85-95% (depends on dataset complexity) [file:28] MNIST: Test accuracy typically 97%+ with basic models [file:29]

Example Outputs

Decision boundary visualization and tree structure plots in notebooks. MNIST predictions show correctly classified digits with confidence scores. [file:28][file:29]

Requirements

• jupyter==1.0.0
• pandas==2.2.3
• numpy==2.1.1
• scikit-learn==1.5.2
• matplotlib==3.9.2
• seaborn==0.13.2
• pillow==10.4.0

Related Resources

MNIST Dataset - Full TMNST Dataset at - https://www.kaggle.com/datasets/hojjatk/mnist-dataset Scikit-learn Decision Trees