Metode Bidirectional Associative Memory (BAM) Kontinu Pengenalan Pola Karakter Untuk Keamanan Data

Abstract

Artificial Neural Networks have a paradigm for processing information with the concept of working using biology in processing information similar to how the human brain works. Artificial neural networks solve many problems using the concept of uncertainty such as the discussion in the material, namely the introduction of Continuous BAM Character Patterns. The problem in this research is the lack of data security in maintaining information so that a lot of data can be accessed by people who do not have authority. The main objective of this research is to maintain data security using the Continuous BAM concept. The BAM Kontine method is a method that has the ability to have associative memory or content addressable memory that can be called by using the part stored in the memory itself. The Continuous BAM method will change input to output more smoothly with values that lie in the range [0,1]. The activation function used is the sigmoid function. The results obtained from x1 = [7,-11], x2= [7,13] and x3=[-1,9], All Signs can be recognized by the Continuous BAM algorithm. All Sign Patterns have the same target but have different values.