Neural Network with Tanh Activation

Overview

This project implements a simple neural network with one hidden layer using Tanh activation function. The network consists of:

• Input Layer: 2 neurons
• Hidden Layer: 2 neurons
• Output Layer: 1 neuron

It initializes random weights, performs forward propagation, computes the final output, and calculates the error between the predicted and target values.

Features

• Uses Tanh Activation Function.
• Random weight initialization.
• Computes Mean Squared Error (MSE).
• Visualizes the Tanh function using matplotlib.

Dependencies

Make sure you have Python and the following libraries installed:

pip install numpy matplotlib

How It Works

1. Initializes weights randomly within the range (-0.5, 0.5).
2. Computes activations for the hidden and output layers using tanh.
3. Computes error using Mean Squared Error (MSE) formula.
4. Plots the Tanh activation function for visualization.

Code Structure

• tanh_activation(x): Computes the Tanh activation.
• initialize_weights(shape): Initializes random weights.
• compute_error(target, output): Computes the error.
• Forward pass:
  • Computes hidden layer activations.
  • Computes final output.
• Prints results and plots the Tanh Activation Function.

Example Output

Hidden Layer Input: [ 0.2 -0.1]
Hidden Layer Output: [ 0.197 -0.099]
Output Layer Input: [0.45]
Final Output: [0.42]
Error: [0.0035]

Visualization

The script generates a plot of the Tanh Activation Function to illustrate its shape and properties.

License

This project is open-source. Feel free to modify and use it for educational purposes.