Implementasi Jaringan Syaraf Tiruan Menggunakan Metode Heteroassociative Memory Pada Kode Rahasia Menggunakan System Biner

Abstract

Letter pattern recognition can be completed with artificial neural networks using the Heteroassociative Memory method. Pattern recognition can be completed using a complex system so as to produce an information that can be used as a system to determine the confidentiality of a data. Heteroassociative Memory is capable of storing various kinds of collections based on data grouping, with each weight determined. The vector of belongs to each data group. The case solved in this research is determining the password to find out the input results match the predetermined target. The known password is a binary number password which consists of 6 data conversions and has targets T1 and T2. From the 6 initial weight initialization data inputs (w), the values w11=3, w21=3, w31=1, w41=2, w12=0, w22=0, w32=0, w42=0 are obtained. After testing with the training input, all activation function searches have values according to predetermined targets, so that it can be concluded that all patterns can be recognized by the system.