🩷Breast Cancer Prediction using Machine Learning

📌 Project Overview

This project aims to develop a machine learning model that predicts whether a tumor is malignant or benign based on medical features. The dataset used is the Breast Cancer Wisconsin Dataset from sklearn.datasets.

🛠️ Steps in the Project

1️⃣ Data Loading & Preparation

• Loaded the Breast Cancer Wisconsin Dataset and converted it into a pandas.DataFrame.
• Separated features (x) and the target variable (y).

2️⃣ Exploratory Data Analysis (EDA)

• Checked dataset dimensions (569 samples, 30 features).
• Verified missing values (none found).
• Used describe() to analyze key statistics such as mean and standard deviation.

3️⃣ Splitting Data into Training & Testing Sets

• Used train_test_split() to split data into 80% training and 20% testing.

4️⃣ Training a Random Forest Classifier

• Trained an initial Random Forest Model (RandomForestClassifier) with default hyperparameters.
• Initial Accuracy: 96.49%

5️⃣ Comparing Different Models

• Tested Logistic Regression and SVC (Support Vector Classifier) alongside Random Forest.
• Results:
  • Random Forest: 96.49%
  • Logistic Regression: 95.61%
  • SVC: 94.74%

6️⃣ Feature Importance Analysis

• Identified key features contributing to the prediction.
• Most important features:
  • Concave points (worst)
  • Area (worst)
  • Radius (worst)
  • Concave points (mean)

7️⃣ Hyperparameter Tuning with GridSearchCV

• Used GridSearchCV to find the best Random Forest parameters:
  • n_estimators = 150
  • max_depth = None
• Optimized Accuracy: 96.26%

8️⃣ Code Optimization with Classes & Functions

• Refactored the code into a structured class (BreastCancerPrediction) for better readability and maintainability.

---

📌 Key Results

✅ The trained Random Forest model achieves an accuracy of ~96%, making it highly reliable for breast cancer classification.
✅ Feature Importance Analysis helped identify the most relevant medical features.
✅ Hyperparameter Tuning slightly improved the model’s performance.

---

🚀 Future Enhancements

• User Input Feature: Allow users to enter their own data and get predictions.
• Further Model Optimization: Try deep learning (e.g., neural networks) for comparison.
• Apply to Other Medical Datasets to test generalization.

---

📂 Installation & Usage

To run this project locally, follow these steps:

🔹 Install dependencies

pip install pandas scikit-learn matplotlib

🔹 Run the script

python breast_cancer_prediction.py

Dataset Information

The dataset used in this project is the Breast Cancer Wisconsin Dataset, available in sklearn.datasets. It contains:

• 569 samples
• 30 numerical features
• Binary target variable (0 = malignant, 1 = benign)

For more details, visit the Breast Cancer Dataset Documentation

🤝 Contributing

Feel free to contribute by:

• Improving the model
• Adding a web interface
• Exploring different machine learning techniques

⭐ If you find this project useful, consider giving it a star on GitHub!