Handwritten Digit Recognition

This project demonstrates handwritten digit recognition using multiple traditional machine learning models applied to the MNIST dataset. We evaluated and compared the performance of each model and visualized results using confusion matrices and accuracy charts.

---

Setup Instructions

1. Clone the Repository

git clone https://github.com/Vansh-Agrawal-IIT-J-27/Handwritten_Digit_Recognition.git
cd Handwritten_Digit_Recognition

2. Install Requirements

pip install -r requirements.txt

3. Run the Main Script

python main.py

All confusion matrices and model accuracy summaries will be saved automatically.

---

Models Implemented

• K-Nearest Neighbors (KNN)
• Support Vector Machine (SVM)
• Decision Tree
• PCA + KNN
• Naive Bayes
• Logistic Regression
• Random Forest
• Linear Discriminant Analysis (LDA)

Each model is evaluated using classification reports and confusion matrices.

---

Evaluation Outputs

File	Description
results.txt	Accuracy of all models
confusion_matrices/*.png	Confusion matrices for each model
model_accuracy_comparison.png	Visual bar chart of model accuracies

---

Team Members & Contributions

S. No.	Name	Contribution
1.	Vansh Agrawal	Code architecture, model training, GitHub
2.	Chandavath Akhil	Data preprocessing, SVM & KNN models
3.	Sapavath Gharulal	Logistic Regression, accuracy evaluation
4.	Jatavath Sudheer	Decision Tree, Random Forest models
5.	Banoth Mallesh	LDA, Naive Bayes implementation
6.	Golla Sathvik	Confusion matrix visualization, summary chart

---

Folder Structure

Handwritten_Digit_Recognition/
│
├── main.py
├── results.txt
├── requirements.txt
├── README.md
├── model_accuracy_comparison.png
│
├── data/
│   └── mnist.npz
│
├── models/
│   ├── knn_model.py
│   ├── svm_model.py
│   ├── tree_model.py
│   ├── pca_knn_model.py
│   ├── naive_bayes_model.py
│   ├── logistic_regression_model.py
│   ├── random_forest_model.py
│   └── lda_model.py
│
├── utils/
│   ├── evaluation.py
│   └── plot_summary.py
│
└── confusion_matrices/
    ├── knn.png
    ├── svm.png
    ├── tree.png
    ├── pca_knn.png
    ├── nb.png
    ├── lr.png
    ├── rf.png
    └── lda.png

---

License

This project is for academic and learning purposes only.