🔢 Digit Pattern Classifier

A simple, straightforward machine learning pipeline designed to recognize handwritten digits (0-9) from 8x8 pixel images.

This project explores traditional machine learning algorithms and compares their performance using both raw pixel data and dimensionally reduced data (PCA).

This project uses the optical recognition of handwritten digits dataset (a copy of the UCI ML hand-written digits dataset) built into scikit-learn.

• Classes: 10 (Digits 0 through 9)
• Samples: 1,797 images
• Features: 64 (8x8 grid of integer pixels)

The following classifiers are trained and evaluated based on Accuracy, Precision, and Recall:

1. Random Forest Classifier
2. Support Vector Machine (SVM) (Polynomial Kernel)
3. Artificial Neural Network (ANN) (Multi-Layer Perceptron)
4. K-Nearest Neighbors (KNN)

• 01_baseline_classification.ipynb: Trains the four models on the normalized 64-pixel feature set.
• 02_pca_classification.ipynb: Applies Principal Component Analysis (PCA) to reduce the feature space from 64 to 32 dimensions before training, demonstrating the trade-off between dimensionality reduction and accuracy.

1. Clone the repository

git clone [https://github.com/yourusername/Digit-Pattern-Classifier.git](https://github.com/yourusername/Digit-Pattern-Classifier.git)
cd Digit-Pattern-Classifier