Deployment Summary: Breast Cancer Prediction Web App

Project Title: Breast Cancer Tumor Classifier Model: Random Forest Classifier using top 15 features Dataset: Breast Cancer Wisconsin Diagnostic Dataset (sklearn.datasets.load_breast_cancer) Deployment Type: Web Application (HTML + Flask) + API Hosting Platform: Render (Free Cloud Hosting)

---

Steps Followed

1. Data Preparation:

   • Loaded the breast cancer dataset from scikit-learn.
   • Selected the top 15 most relevant features using SelectKBest with f_classif.

2. Model Building:

   • Trained a compact RandomForestClassifier with n_estimators=25, max_depth=6.
   • Evaluated accuracy on test data (approx. ~94–96%).
   • Saved the model as model.joblib using joblib.dump().

3. Web App Development:

   • Created a Flask app (app.py) to load the saved model and take 15 inputs via a form.
   • Designed a user-friendly HTML form (index.html) for the input interface.
   • Rendered prediction output directly on the webpage (malignant or benign).

4. API Integration:

   • Exposed a POST endpoint /predict to receive form data and return prediction results dynamically.

5. Deployment:

   • Used Render.com (free tier) to deploy the web app.
   • Created requirements.txt, Procfile, and pushed the project to GitHub.
   • Linked GitHub repository to Render for auto-deployment.

---

Outcome

• A live, browser-accessible breast cancer classifier that accepts 15 feature values and predicts whether a tumor is malignant or benign.
• Fully functional on free cloud infrastructure (Render).
• Project includes both web form interface and API logic.