K-Nearest Neighbors (KNN) Classification Project

Objective

Implement KNN classifier, experiment with different K values, evaluate model, and visualize decision boundaries.

Dataset

• Place your CSV dataset inside the data/ folder.
• By default, if no dataset is provided, the script uses the Iris dataset from scikit-learn.

Features

• Train/Test split
• Data normalization
• KNN training with multiple K values
• Accuracy and confusion matrix evaluation
• Decision boundary visualization (2D)
• Best K selection

Folder Structure

knn_classification_project/
│
├── data/                 # Place dataset.csv here
├── outputs/              # Generated plots & metrics
├── src/                  # Python code
│   └── knn_model.py
├── README.md
└── requirements.txt

How to Run

1. Create virtual environment (optional but recommended):

python -m venv venv 
venv\Scripts\activate   

2. Install dependencies:

pip install -r requirements.txt

3. Run script (default dataset):

python src/knn_model.py

4. Run script with your own dataset:

python src/knn_model.py --csv data/your_dataset.csv --target target_column_name --features col1 col2

Note: Decision boundary visualization works only if you specify exactly 2 features.

Outputs

• outputs/accuracy_vs_k.png
• outputs/confusion_matrix.png
• outputs/decision_boundary.png (if 2D features)
• outputs/metrics.json