Hybrid conjugate gradient as neural network using new line search technique
Other Title(s)
التدرج المترافق كشبكة عصبية هجينة باستخدام خط بحث جديد
Author
Source
College of Basic Education Researches Journal
Issue
Vol. 5, Issue 3 (30 Sep. 2007), pp.214-233, 20 p.
Publisher
University of Mosul College of Basic Education
Publication Date
2007-09-30
Country of Publication
Iraq
No. of Pages
20
Main Subjects
Information Technology and Computer Science
Topics
Abstract AR
في هذا البحث تم استحداث أسلوب خوارزمية التدرج المترافق كشبكة عصبية و المعتمدة على خوارزمية وسطية يتم بموجبها حساب نسبة التعلم الملائمة لهذه الشبكة و بالتالي يتم الحصول على شبكة عصبية جديدة ذات سرعة عالية و من نوع Supervised. و أن هذا الأسلوب الرياضي المقترح كشبكة عصبية جديدة قد أثبت كفاءة عالية من حيث الحسابات المعتمدة و لإدخالات مختلفة من الحروف الإنكليزية.
Abstract EN
In this paper a conjugate gradient algorithm was devised as a neural network, based on a middle algorithm, according to which the appropriate learning rate for this network was calculated.
Accordingly, a new neural network with high speed and of a supervised type was obtained.
This mathematical style proposed as a new neural network proved to be efficient with regard the results for different English letters.
American Psychological Association (APA)
Muhammad, Shadha Abd Allah. 2007. Hybrid conjugate gradient as neural network using new line search technique. College of Basic Education Researches Journal،Vol. 5, no. 3, pp.214-233.
https://search.emarefa.net/detail/BIM-339438
Modern Language Association (MLA)
Muhammad, Shadha Abd Allah. Hybrid conjugate gradient as neural network using new line search technique. College of Basic Education Researches Journal Vol. 5, no. 3 (2007), pp.214-233.
https://search.emarefa.net/detail/BIM-339438
American Medical Association (AMA)
Muhammad, Shadha Abd Allah. Hybrid conjugate gradient as neural network using new line search technique. College of Basic Education Researches Journal. 2007. Vol. 5, no. 3, pp.214-233.
https://search.emarefa.net/detail/BIM-339438
Data Type
Journal Articles
Language
English
Notes
Includes appendix : p. 233
Record ID
BIM-339438