DeepWater Predictor 🌊

DeepWater Predictor is a Deep Learning project designed to predict whether a plant needs watering based on environmental factors. By leveraging a Artificial Neural Network built with TensorFlow and Keras, this project demonstrates how deep learning can be applied to precision agriculture and smart home automation.

🚀 Overview

This project was developed as part of a Deep Learning One-Shot learning journey. It focuses on the fundamental concepts of neural networks, including layer stacking, activation functions, and gradient descent optimization. The model analyzes environmental variables to make intelligent predictions, helping optimize water usage for plant care.

🧠 Deep Learning vs Machine Learning

While this project falls under the umbrella of Machine Learning, it specifically uses Deep Learning techniques:

• Neural Networks: Unlike traditional ML (like Decision Trees or SVMs), this project uses layers of interconnected "neurons".
• Feature Learning: The model learns complex relationships between soil moisture, temperature, and sunlight through hidden layers.
• Frameworks: Utilizes specialized DL libraries like TensorFlow and Keras.

📊 Features

• Sequential Neural Network: Implementation of a multi-layer perceptron for binary classification.
• Normalization Pipeline: Custom feature scaling to ensure stable and faster model convergence.
• Training Variants: Includes implementations for:
  • Full-Batch Gradient Descent
  • Mini-Batch Gradient Descent
  • Stochastic Gradient Descent (SGD)

🛠️ Tech Stack

• Core: Python 3.x
• Deep Learning: TensorFlow, Keras
• Data Science: Pandas, NumPy
• Preprocessing: Scikit-learn

📉 Dataset Schema

The model is trained on synthetic environmental data:

Feature	Description
Soil Moisture	Normalized level of moisture in the soil
Temperature (C)	Ambient temperature in Celsius
Sunlight Hours	Daily exposure to sunlight
Needs Water	Target label (1: Yes, 0: No)

🏗️ Model Architecture

The neural network is structured as follows:

1. Input Layer: 3 features (Moisture, Temperature, Sunlight).
2. Hidden Layer: 8 neurons with ReLU activation for non-linear mapping.
3. Output Layer: 1 neuron with Sigmoid activation (mapping output to 0-1 probability).

⚙️ Installation & Usage

1. Clone the repository:

   git clone https://github.com/rohansachinpatil/DeepWater_Predictor.git

2. Install dependencies:

   pip install tensorflow pandas numpy scikit-learn

3. Run the Notebook: Open DeepWater_Predictor.ipynb in your preferred environment (Jupyter, VS Code, or Google Colab).

🤝 Contributing

Contributions, issues, and feature requests are welcome! Feel free to fork and submit a PR.

📜 License

This project is licensed under the MIT License.