PENGENALAN TULISAN TANGAN MENGGUNAKAN JARINGAN SYARAF TIRUAN DENGAN ALGORITMA PERCEPTRON

Abstract

Handwriting is a product of an individual's writing activity. The unique characteristics of handwriting vary from person to person, reflecting their individual traits. However, handwriting is often misused by irresponsible parties. Currently, there is a lack of tools to facilitate the recognition or verification of handwriting. Generally, handwriting recognition can be addressed using Artificial Neural Networks (ANNs). ANNs are artificial representations of the human brain that attempt to simulate the brain's learning processes. In this study, the Perceptron algorithm, a supervised learning method within neural networks, is employed. The research uses 5x5 pixel images of handwriting samples, with the Perceptron algorithm functioning based on the weight values obtained. Testing with the Perceptron algorithm showed that it can recognize three different handwriting samples. The testing concluded at Epoch 3, where the value of Y_in = 39 exceeded the threshold value of 30.