🌾 DON Concentration Prediction App

A machine learning-powered web application for predicting Deoxynivalenol (DON) concentration in agricultural products using spectral data. Built with Python, Streamlit, FastAPI, and Scikit-Learn, this project enables quick and accurate DON contamination predictions.

🚀 Features

✅ User-friendly Interface – Enter spectral values and get predictions instantly ✅ Machine Learning Model – Uses a trained MLP Regressor (Neural Network) ✅ Data Preprocessing – Handles missing values, normalizes spectral features ✅ Model Evaluation – Performance measured using MAE, RMSE, and R² ✅ Web Deployment – Available via Streamlit (UI) & FastAPI (API)

📂 Project Structure

├── mycotoxin_prediction.py # Main ML pipeline: data processing, training, evaluation ├── app.py # FastAPI backend for model inference ├── don_prediction_model.pkl # Trained model ├── scaler.pkl # StandardScaler for input normalization ├── MLE-Assignment.csv # Dataset ├── README.md # Project documentation

🛠️ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/your-username/don-prediction-app.git cd don-prediction-app

2️⃣ Install Dependencies

  pip install -r requirements.txt

3️⃣ Run the Streamlit App

 streamlit run mycotoxin_prediction.py

4️⃣ Run the FastAPI Backend

uvicorn app:app --reload

📊 Model Performance

Metric Score Mean Absolute Error (MAE): 3342.7450 Root Mean Squared Error (RMSE): 9834.3624 R² Score: 0.6540 write this as a table

📢 API Usage

Once the FastAPI backend is running, send a POST request to: http://localhost:8501/predict/

Example Request (JSON)

{ "features": [0.1, 0.5, 0.3, 0.8, 0.9,] }

Example Response

{ "predicted_don_concentration": 3.45 }

🏗️ Future Improvements

🔍 Hyperparameter tuning for improved model accuracy 📈 Visualization dashboards for better insights ☁️ Cloud deployment on AWS/GCP

🖥️ Author

👩‍💻 Sudarshanam Yessasvini 📧 Email: your.yessasvini.s@gmail.com

🔗 Deployment URLs

Local URL: http://localhost:8501 Network URL: http://192.168.1.2:8501 (Change this according to your network configuration)