Fish Disease Detection from Water quality parameters using SVM 🐟

A machine learning-based Streamlit web application for predicting Tilapia fish disease risk using IoT water quality parameters like Dissolved Oxygen and pH. This app supports training, evaluating, saving, and making real-time predictions.

---

💥 Features

• 📤 Upload Excel dataset with timestamped IoT water quality data
• 🔍 Train an SVM classifier using grid search with hyperparameter tuning
• 📈 Visualize model metrics: Accuracy, Precision, Recall, F1-score
• 🔥 Confusion matrix and decision boundary plots
• 💾 Save and load trained models --

📊 Dataset Assumptions

Your Excel file should contain the following columns:

• Datetime
• Dissolved Oxygen (mg/L)
• pH
• Disease Occurrence (Cases)

A binary target disease is derived internally as:

df['disease'] = (df['Disease Occurrence (Cases)'] > 1.5).astype(int)

📦 Installation

Clone the repository

git clone https://github.com/your-username/fish-disease-prediction.git
cd fish-disease-prediction

Create a virtual environment

python -m venv venv
source venv/bin/activate      # For Linux/macOS
venv\Scripts\activate         # For Windows

Install dependencies

pip install -r requirements.txt

Run the Streamlit app

streamlit run frontend.py

📁 Project Structure

.
├── core.py               # Core ML logic (data processing, training, evaluation)
├── frontend.py           # Streamlit frontend UI
├── model.pkl             # Saved trained model (generated after training)
├── requirements.txt      # Python dependencies
└── README.md             # Project documentation

---

💡 How to Use

1. Upload your IoT dataset (Excel) via the sidebar

2. Configure test window size and SVM parameters (C, gamma, kernel)

3. Click “Run Model” to train and evaluate

4. Visualize performance and decision boundaries

5. Click “Save Model” to persist the best model

6. Switch to the “Predict” tab to use manual inputs for disease prediction