MalayaKumar6589/Lung-Cancer-Detection

Disease Prediction Web Application

This repository contains the source code for a Machine Learning-powered web application that predicts the likelihood of a patient having a specific disease based on their medical history and lifestyle factors.

The backend is built with FastAPI and serves a frontend using Jinja2 templates. It processes user medical data, performs feature engineering on the fly, and serves predictions using a pre-trained ensemble model (votingModel.pkl).

🌐 Live Deployment: This application is deployed and hosted live on Render.

🚀 Features

• —FastAPI Backend: A fast, modern, and asynchronous Python web framework handling routing and model inference.
• —Machine Learning Integration: Uses a pre-trained Voting Classifier model to predict the presence of disease (Has Disease vs No Disease).
• —Robust Feature Engineering: Automatically calculates derived medical indicators directly from the form inputs, including:
• —smoking_cancer_factor
• —metabollic_risk
• —adjusted_survival_index
• —Data Categorization: Converts raw continuous inputs (BMI, cholesterol, age, dates) into standard categorical buckets for the ML model using custom helper functions.
• —Categorical Encoding: Uses pre-fitted Ordinal Encoders to reliably transform categorical text inputs (e.g., country, gender, treatment type) into numerical features.

🛠️ Tech Stack

• —Backend Framework: Python 3, FastAPI, Uvicorn
• —Data Processing: Pandas
• —Machine Learning: Scikit-Learn (Pickle for model/encoder serialization)
• —Frontend: HTML/CSS (served via Jinja2Templates & StaticFiles)
• —Deployment: Render

📁 Project Structure

• —`main.py`: The main entry point of the application. It initializes the FastAPI app, configures the template and static directories, handles the / and /predict routes, and houses the core data pipeline for model inference.
• —`helper.py`: Contains data processing utilities to categorize continuous variables (e.g., categorizing BMI into weight classes, Cholesterol into risk levels, and calculating age gaps between diagnosis and treatment).
• —`encoder.py`: Responsible for loading six distinct pre-trained ordinal encoders (ord1.pkl through ord6.pkl) and transforming categorical form data into numerical arrays.
• —`templates/`: Directory containing the index.html file (the user interface).
• —`static/`: Directory for static assets (CSS, JS, images).
• —*`.pkl` files**: Serialized Machine Learning artifacts (The Voting Model and 6 Ordinal Encoders).

⚙️ Data Pipeline

When a user submits their data via the /predict endpoint, the application performs the following steps:

1. 1.Type Casting: Converts form strings to numeric values (age, bmi, hypertension, etc.).
2. 2.Categorization (`helper.py`): Bins bmi, cholesterol_level, and age into predefined risk/stage tiers. Calculates end_treatment_age based on diagnosis and treatment dates.
3. 3.Encoding (`encoder.py`): Transforms categorical inputs into machine-readable numerical formats using the loaded pickle encoders.
4. 4.Feature Generation: Calculates the metabolic risk, smoking cancer factor, and adjusted survival index based on mathematical relationships between the inputs.
5. 5.Inference: Drops intermediate columns, orders the 14 final features exactly as the model expects, and generates a prediction.

💻 Local Setup & Installation

If you wish to run this project locally on your machine, follow these steps:

1. Prerequisites

Ensure you have Python 3.8+ installed. You will also need Git. (Note: Because this project contains large `.pkl` files, ensure you have Git LFS installed to pull the models correctly).

2. Clone the Repository

git clone [https://github.com/yourusername/your-repo-name.git](https://github.com/yourusername/your-repo-name.git)
cd your-repo-name

3. Create a Virtual Environment (Recommended)

python -m venv venv
# On Windows:
venv\Scripts\activate
# On Mac/Linux:
source venv/bin/activate 

4. Install Dependencies

pip install fastapi uvicorn pandas scikit-learn jinja2 python-multipart

5. Run the Development Server

Start the Uvicorn server by running main.py directly:

python main.py

Alternatively, run it via uvicorn:

uvicorn main:app --reload

6. Access the App

Open your web browser and navigate to: http://localhost:8000