🫁 Lung Cancer Prediction

This project combines machine learning with a Flask web interface to predict the risk of lung cancer and survival likelihood based on user input. It consists of two trained models: one for lung cancer risk prediction, and another for survival analysis using comorbidities and clinical indicators.

⚠️ Experimental Project Notice
This tool is created for educational and experimental purposes only. It is not a medical device. The results are derived from limited, small-scale datasets and do not represent clinically validated predictions.

---

⚙️ Functionality

• Lung Cancer Risk Prediction
  Uses a Random Forest model trained on basic survey features like smoking, fatigue, anxiety, etc.

• Survival Prediction
  Uses a Decision Tree model trained on synthetic clinical data including comorbidities, tumor stage, and survival duration classes.

---

🧪 Limitations

• Small and potentially biased datasets
• Simplified feature encoding
• No hyperparameter tuning or cross-validation
• Not suitable for clinical decision-making

---

🛠️ Tech Stack

• Python 3.x
• Flask
• pandas, numpy
• scikit-learn
• joblib
• matplotlib, seaborn

---

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/anilcemelemir/LungCancerPrediction.git
cd LungCancerPrediction

2. Install Dependencies

pip install -r requirements.txt

3. Train the Models (Once)

python src/train_cancer_model.py
python src/train_survival_model.py

Make sure both .csv files are inside the data/ folder before training.

4. Run the Flask App

python main.py

Go to http://localhost:5000 in your browser.

---

📄 License

MIT License © 2025 Semih Çetin, Velihan Özge, Anıl Cem Elemir

---

🏷️ Tags

machine-learning
flask
lung-cancer
healthcare
survival-analysis
classification
python
experimental