👥 K-Nearest Neighbors (KNN) — Classification Project

A supervised machine learning project that applies the K-Nearest Neighbors (KNN) algorithm to classify data points based on feature similarity and distance metrics.

This project demonstrates distance-based classification, feature scaling, model selection, and evaluation.

---

📌 Overview

This project builds a KNN classifier using a labeled dataset and evaluates how prediction accuracy changes with different values of K (number of neighbors).

The workflow includes preprocessing, scaling, model training, hyperparameter selection, and performance evaluation.

---

🎯 Objective

Predict the class label of observations by:

• Measuring distance to nearest neighbors
• Voting based on closest K samples
• Selecting the optimal K value

---

🧠 Learning Type

Supervised Learning — Classification

• Uses labeled training data
• Distance-based prediction
• Instance-based learning method

---

🧩 Tech Stack

• Python
• Pandas
• NumPy
• Scikit-learn
• Matplotlib
• Seaborn
• Jupyter Notebook

---

📊 Dataset

The dataset contains numeric feature variables and a target class label.

Typical steps included:

• Feature inspection
• Class distribution check
• Train/test split

(Target column name depends on dataset used.)

---

🔬 Project Workflow

1️⃣ Data Exploration

• Dataset inspection
• Feature distribution analysis
• Correlation checks

---

2️⃣ Preprocessing

• Train/test split
• Feature scaling (critical for KNN)
• Standardization using scaler

---

3️⃣ Model Training

KNN classifier used: sklearn.neighbors.KNeighborsClassifier

Model trained with multiple K values.

---

4️⃣ Choosing Optimal K

• Error rate calculated across K range
• Error vs K plotted
• Best K selected based on lowest error

---

5️⃣ Evaluation

Model evaluated using:

• Accuracy
• Confusion Matrix
• Classification Report
• Error rate comparison

---

📈 Results

              precision    recall  f1-score   support

           0       0.73      0.72      0.73       148
           1       0.73      0.74      0.74       152

    accuracy                           0.73       300
   macro avg       0.73      0.73      0.73       300
weighted avg       0.73      0.73      0.73       300

---

⚙️ Key Concepts Demonstrated

• KNN algorithm
• Distance-based learning
• Feature scaling importance
• Hyperparameter tuning (K selection)
• Bias vs variance tradeoff
• Classification metrics

---

▶️ How to Run

Clone repository

git clone https://github.com/rohitb281/knn-project.git
cd knn-project

Install dependencies

pip install pandas numpy scikit-learn matplotlib seaborn jupyter

Launch notebook

jupyter notebook

Open:

K Nearest Neighbors Project.ipynb

• Run all cells.

---

#@ 🚀 Possible Improvements

• Distance metric comparison
• Cross-validation for K selection
• Weighted KNN
• Feature selection
• Pipeline automation

---

⚠️ Limitations

• Sensitive to feature scaling
• Slower prediction on large datasets
• Performance depends on K choice

---

📄 License

• Open for educational and portfolio use.

---

👤 Author

• Rohit Bollapragada
• itHub: https://github.com/rohitb281