MNIST Binary Classifier

This project implements a simple binary classifier (perceptron) to recognize the digit "2" from the MNIST dataset. It includes custom data loading, training, and evaluation functionalities in C#, without reliance on external machine learning libraries.

Project Structure

• MNISTLoader: Handles loading and processing the MNIST dataset from binary files.
• BinaryClassifier: A single-layer perceptron model for binary classification.
• Program.cs: Executes training and evaluation, printing model accuracy.

Setup Instructions

Download MNIST Data:

Place train-images-idx3-ubyte, train-labels-idx1-ubyte, t10k-images-idx3-ubyte, and t10k-labels-idx1-ubyte files in the assets folder with subfolders like below:

assets/
├── train-images-idx3-ubyte/
│   └── train-images.idx3-ubyte
├── train-labels-idx1-ubyte/
│   └── train-labels.idx1-ubyte
├── t10k-images-idx3-ubyte/
│   └── t10k-images.idx3-ubyte
└── t10k-labels-idx1-ubyte/
    └── t10k-labels.idx1-ubyte

Run the Project:

The project will load the MNIST data, train the binary classifier over 10 epochs, and evaluate its accuracy in distinguishing the digit "2" from other digits.

Code Overview

MNISTLoader

• LoadMNIST: Loads and processes images and labels from binary files.
• LoadImages and LoadLabels: Handle loading images and labels into double and integer arrays for model processing. Automatically adjusts for dataset size (60,000 for training, 10,000 for testing).

BinaryClassifier

• Train: Trains the perceptron model on labeled MNIST images to recognize the digit "2".
• Evaluate: Calculates the model's accuracy on the test dataset.
• Sample Execution Output After running the program, you should see output similar to:

Epoch 1/10 completed.
Epoch 2/10 completed.
...
Epoch 10/10 completed.
Accuracy: 96.51%

Customizing

• Epochs: Modify the number of epochs in Program.cs by adjusting classifier.Train(..., epochs: 10);.
• Learning Rate: Adjust the learning rate in BinaryClassifier for more fine-tuned training.