Pengenalan Angka Tulisan Tangan Menggunakan Jaringan Syaraf Buatan

Abstract

Background: It is a chalange to build a program that
behave intelligent like human. Specially how to build a
program that able to learn in order to solve a specific
problem. In this reserarch will be developed a computer
program implementing an artificial neural network that
learns to recognize handwritten digits. The focus on
handwritting recognition because it is an excellent
prototype problem for learning about neural network in
general.
Aims: Building an intelligent program that able to
recognize hand written number (digits).
Methode: Artificial Neural Network with
Backpropagation architecture and Stochastic gradient
descent learning algorithm wich is implemented in Python
programming language.
Result: The program can recognize handwritten digits
with an accuracy over 96 percent, without human
intervention.
Conclusion and Advice: A good learning algorithm that
be learned with bad learning data will perform worse than
a simple learning algorithm that be learned with good
learning data. It is suggested that this research to be
improved with capability to recognize handwritten letter.