RPropNetBuilder is an implementation of a dynamic multilayer neural network. It simulates both forward and backward propagation, allowing for adjustments to network parameters such as the number of internal layers and nodes.

• Forward and backward propagation simulation
• Adjustable internal layers and nodes
• Utilizes the RProp weight update algorithm
• Supports activation functions (Identity, Sigmoid) and error functions (Sum of Squares, Cross Entropy)

It's a Python-based program for creating and training neural networks, specifically designed to solve a 10-class classification problem using the MNIST dataset.

• Python
• Neural network framework (implemented from scratch)

The project is in a complete state, with all functionalities implemented and tested using the MNIST dataset. Further development may include extending capabilities or adapting it for different datasets or problems.

The project aims to demonstrate proficiency in neural network implementation and training. It's designed to solve a specific classification problem using established algorithms.