🧠 Binary Classification with PyTorch

This repository contains a simple neural network implementation in PyTorch for performing binary classification on synthetic data generated using scikit-learn.

📌 Overview

The goal is to train a fully connected neural network to classify samples into two classes using synthetic data. This is a basic example that demonstrates how to:

• Generate classification data
• Define a feedforward neural network
• Train and evaluate a PyTorch model
• Visualize training loss

📊 Dataset

The dataset is generated using make_classification from sklearn.datasets with the following parameters:

• n_samples=1000
• n_features=10
• n_classes=2

Data is then split into training (70%) and test (30%) sets.

🧠 Model Architecture

The model is a simple feedforward neural network with:

• Input layer: 10 features
• Hidden layer: 64 neurons, ReLU activation
• Output layer: 2 classes (for binary classification)

class NeuralNetwork(nn.Module):
    def __init__(self, input_size):
        super(NeuralNetwork, self).__init__()
        self.fc1 = nn.Linear(input_size, 64)
        self.relu = nn.ReLU()
        self.fc2 = nn.Linear(64, 2)

    def forward(self, x):
        x = self.relu(self.fc1(x))
        x = self.fc2(x)
        return x

⚙️ Training

• Loss Function: CrossEntropyLoss
• Optimizer: Adam with learning rate 0.01
• Epochs: 20

Training loss is recorded and printed at each epoch.

✅ Evaluation

The model is evaluated on the test set using accuracy_score from scikit-learn.

Example output:

Accuracy: 88.67%

📉 Training Loss Plot

A line plot is generated showing the training loss across epochs.

(Save the figure using plt.savefig("training_loss.png") if you want to include the image in GitHub)

📂 File Structure

├── classification_nn.py       # Main training script
└── README.md                  # Project documentation

▶️ How to Run

Make sure you have the required packages:

pip install torch matplotlib scikit-learn

Then run the script:

python classification_nn.py