BREAST_CANCER_PREDICTION

Breast Cancer prediction project using basic neural network

---

🧬 Breast Cancer Prediction using Neural Networks

This project showcases a binary classification model built using a custom Neural Network in PyTorch to predict whether a tumor is Benign (0) or Malignant (1). The dataset is derived from diagnostic measurements in medical reports.

---

📁 Dataset

The dataset is a tabular CSV file with numeric features representing medical metrics extracted from cell nuclei images.

🔹 Structure:

• Feature Columns: Continuous variables (e.g., radius, texture)
• Target Column: Diagnosis with labels 'B' (Benign) and 'M' (Malignant)

🏷️ Class Labels

{'B': 0, 'M': 1}

---

🧠 Model Architecture

The neural network is implemented using torch.nn.Module and includes:

• Input Layer: Matches the number of features
• 2 Hidden Layers: Each with ReLU activation
• Output Layer: Single neuron with Sigmoid activation (for binary classification)
• Loss Function: Binary Cross Entropy
• Optimizer: Adam

---

🔧 Hyperparameters

Parameter	Value
Batch Size	32
Learning Rate	0.001
Epochs	50
Optimizer	Adam
Loss Function	BCEWithLogitsLoss

---

🔄 Workflow

📦 Preprocessing

• Dropped unnecessary columns: id, Unnamed: 32
• Converted labels: 'B' → 0, 'M' → 1
• Split data into training and testing sets
• Converted data into PyTorch tensors

🚂 Training

• Model trained over 50 epochs
• Tracked training loss at each epoch

🧪 Evaluation

• Used accuracy_score from sklearn for final performance
• Confusion matrix plotted using seaborn

Accuracy of the model on the test data: ~96%

---

📊 Visualization

Includes:

• Count plot of label distribution
• Heatmap of feature correlations
• Accuracy and loss plots
• Confusion matrix heatmap

Example:

sns.heatmap(confusion_matrix(y_test, y_pred), annot=True)

---

📦 Libraries Used

Library	Purpose
pandas	Data loading and manipulation
numpy	Numerical operations
matplotlib	Data visualization
seaborn	Statistical plots and heatmaps
torch	Building and training the neural network
sklearn	Metrics and data preprocessing

---

---

▶️ How to Run

1. Clone or download the repository
2. Place data_day1.csv in the project directory
3. Install dependencies:

   pip install pandas numpy matplotlib seaborn torch scikit-learn

4. Open CANCER_PREDICTION_.ipynb in Jupyter Notebook or Google Colab
5. Run all cells sequentially

---

🏆 Results

• Test Accuracy Achieved: ~96%
• Dataset Size: ~569 samples
• High classification performance on medical diagnostic data

---

🙏 Acknowledgements

• PyTorch for model implementation
• Matplotlib & Seaborn for visualization
• Scikit-learn for metrics and preprocessing
• Data inspired by Breast Cancer Wisconsin Dataset

---