Medical Diagnosis Prediction System

🩺 Overview

This project is a medical diagnosis prediction system that predicts possible diseases based on patient-reported symptoms. The web interface is built with Flask and the backend logic leverages Python, Pandas, and machine learning models.

⚙️ Tech Stack

• Python, Pandas, NumPy for data preprocessing
• Pickle for loading trained machine learning models
• Scikit-learn for building the classification model (SVC)
• Flask for serving the web app
• Bootstrap for front-end styling

🛠️ Implementation Steps

1. Prepared datasets for symptoms, precautions, diets, workouts, medications, and disease descriptions
2. Trained an SVC (Support Vector Classifier) to classify diseases based on binary symptom vectors
3. Pickled the model for later use with Flask
4. Created Flask routes:
   • / → main page with symptom input form
   • /predict → processes user symptoms and predicts disease
   • Other static pages: /about, /Contact, /Blog, /Developer
5. On submit, the system returns:
   • Predicted disease
   • Description, precautions, recommended diet, workouts, and medications

📸 UI Previews

Symptom Input Page

Prediction Results Page

Detailed Disease Info Page

📜 Key Code Snippet

@app.route('/predict', methods=['POST'])
def predict():
    symptoms = request.form.get('symptoms')
    user_symptoms = [s.strip() for s in symptoms.split(',')]
    predicted_disease = get_predicted_value(user_symptoms)

    desc, pre, med, die, wrkout = helper(predicted_disease)
    return render_template('index.html',
                           predicted_disease=predicted_disease,
                           dis_des=desc,
                           dis_pre=pre,
                           dis_med=med,
                           dis_diet=die,
                           dis_wrkout=wrkout)
      

This route takes user input symptoms, predicts the disease, and displays all related information.

💡 Lessons Learned

• Preprocessing symptoms into binary vectors enables model predictions
• Pickling models allows fast loading and reduces computation at runtime
• Bootstrap simplifies building a clean and responsive UI quickly
• Handling user input carefully prevents errors in predictions