<![CDATA[Abstrak:. Jaringan Hamming awalnya diteliti oleh Lippmann dan lembaga DARPA, jaringan Hamming bekerja berdasarkan prinsip kecocokan atau kesesuaian data vektor input dengan pola vektor prototip yang tersimpan dalam matriks bobot, sehingga jaringan ini tidak memerlukan “pelatih” dalam proses pembelajaran. Kesederhanaan cara kerja jaringan Hamming membawa konsekuensi akan persyaratan tertentu yang memungkinkan kinerja dalam mendeteksi data vektor input secara akurat. Persyaratan yang diperlukan jaringan Hamming menjadi fokus penelitian pada laporan ini. Variabel data vektor prototip dan variabel data vektor input diteliti untuk melihat pengaruhnya terhadap keakuratan jaringan Hamming dalam mendeteksi data vektor input yang terdapat efek pengaburan dari lingkungan. Hasil penelitian berupa persyaratan yang diperlukan dan selanjutnya dapat diaplikasikan pada sistem Pemanfaatan Jarinngan Saraf Tiruan Hamming dan MAXNET Dalam Mendeteksi Nomor Plat Kendaraan. Kata kunci: Hamming Network, MAXNET, dimensi, keakuratan, bit Abstract: The Hamming network was originally researched by Lippmann and DARPA department, the Hamming network works based on the principle of matching or matching the input vector data with the prototype vector pattern stored in the weight matrix, so this network does not need a "trainer" in the learning process. The simplicity of how the Hamming network works has consequences for certain requirements that enable performance in detecting input vector data accurately. The requirements needed by the Hamming network are the focus of research in this report. Prototype vector data variables and input vector data variables were examined to see their effect on the accuracy of the Hamming network in detecting input vector data that has a defocusing effect from the environment. The results of the research are in the form of requirements that are needed and can then be applied to the system of Utilizing Hamming and MAXNET Artificial Neural Networks in Detecting Vehicle Plate Numbers. Keywords: Hamming Network, MAXNET, Dimension, Acurate, bit]]>
Copyrights © 2022
