Implementasi Algoritma Learning Vector Quantization Untuk Pengenalan Barcode Barang

TL;DR: Implementasi algoritma Learning Vector Quantization untuk pengenalan barcode barang menghasilkan akurasi tinggi dengan data pelatihan yang banyak.

Abstract: Problems in barcode recognition during the barcode identification process. Where when the barcode has noise (damage) then the barcode becomes difficult to recognize. Learning Vector Quantization (LVQ) is a classification method in which each output unit presents a class. LVQ is used for grouping and is also one of the artificial neural networks which is a competitive learning algorithm supervised version of the Kohonen Self-Organizing Map (SOM) algorithm. The purpose of this algorithm is to approach the distribution of vector classes in order to minimize errors in classifying. LVQ learning models are trained significantly to be faster than other algorithms such as the Back Propagation Neural Network. This can summarize or reduce large datasets for a small number of vectors. Based on the results of barcode recognition testing using LVQ algorithm success with training data as much as 4 and conducted calrifikas trial of two data namely: {1,1,1,0} and {1,0,1,1}. Obtained accuracy value generated as much as 90% barcode recognized. The more training data used, LVQ will have a more complete knowledge.