KNN Classification on Breast Cancer Dataset

🎯 Problem Statement and Goal of Project

This project explores a foundational classification task using the K-Nearest Neighbors (KNN) algorithm on the widely used Breast Cancer dataset from scikit-learn. The primary goal is to understand the behavior of KNN in distinguishing between malignant and benign tumors based on numerical features derived from digitized images of breast mass tissue.

---

🛠 Solution Approach

The notebook follows a clean ML workflow:

1. Data Loading: Load breast cancer data via sklearn.datasets.load_breast_cancer().

2. Exploratory Data Analysis:

   • View data shape, features, and class distributions.
   • Use seaborn for basic visualization.

3. Train-Test Split:

   • Split dataset into training and test sets (80/20).

4. Modeling with KNN:

   • Apply KNeighborsClassifier from sklearn.neighbors.
   • Test different k values (number of neighbors).

5. Evaluation:

   • Accuracy evaluation.
   • Confusion matrix plotting.
   • Error rate plot vs. k.

---

🧰 Technologies & Libraries

• Python 3
• numpy, pandas – numerical computing and data manipulation
• seaborn, matplotlib – visualization
• scikit-learn – datasets, model training, evaluation

---

📁 Dataset Description

• Source: sklearn.datasets.load_breast_cancer()

• Features: 30 real-valued features (e.g., mean radius, texture, smoothness)

• Target Classes:

  • 0: malignant
  • 1: benign

• Samples: 569

• Goal: Predict whether a tumor is malignant or benign

---

⚙️ Installation & Execution Guide

1. Download or clone the repository.

2. Install required libraries:

   pip install numpy pandas matplotlib seaborn scikit-learn

3. Run the notebook:

   jupyter notebook knn.ipynb

---

📊 Key Results / Performance

• Train/Test Split: 80/20

• KNN Accuracy: Varies with k, best value selected via error rate plot

• Visualization:

  • Target distribution via seaborn.countplot
  • Confusion matrix for classification performance
  • Error vs. k plot to optimize hyperparameters

---

📸 Sample Outputs

• 📈 Error rate vs. k visualization
• ✅ Confusion matrix
• 📊 Class distribution plots

---

📚 Additional Learnings / Reflections

• Demonstrates clear understanding of:

  • The importance of choosing the right k
  • Effects of class imbalance
  • Train-test split strategy

• A practical and well-scaffolded implementation of KNN in binary classification.

---

👤 Author

mehran Asgari

Email: imehranasgari@gmail.com GitHub: https://github.com/imehranasgari

---

📄 License

This project is licensed under the Apache 2.0 License – see the LICENSE file for details.

---

💡 Some interactive outputs (e.g., plots, widgets) may not display correctly on GitHub. If so, please view this notebook via nbviewer.org for full rendering.