LEADER |
01542nam a22002297a 4500 |
001 |
1758437 |
041 |
|
|
|a eng
|
044 |
|
|
|b السودان
|
100 |
|
|
|9 552956
|a Ali, Omar Balola
|e Author
|
245 |
|
|
|a Over-Fitting Problem Solution with Back-Propagation Neural Network
|
260 |
|
|
|b جامعة أم درمان الإسلامية - كلية التنمية البشرية
|c 2018
|g يناير
|m 1439
|
300 |
|
|
|a 13 - 28
|
336 |
|
|
|a بحوث ومقالات
|b Article
|
520 |
|
|
|b In this paper, we presented three methods for manipulate the over fitting problem, which it done by using the Back Propagation algorithm. Various experiments were done for over fitting problem have discussed. To address over fitting problem, which reflected by the difference between testing and training results, some over fitting solutions have been testing and their results are encouraged. All experiments were done using Handwritten Arabic Characters data set and the performance obtained is robust.
|
653 |
|
|
|a الشبكة العصبية
|a الحروف العربية
|a حل المشكلات
|
692 |
|
|
|b lsolated Handwritten Arabic Character Recognition
|b Back Propagation
|b Feature Extraction
|b Over fitting
|
773 |
|
|
|4 العلوم الإنسانية ، متعددة التخصصات
|6 Humanities, Multidisciplinary
|c 010
|l 004
|m ع4
|o 1432
|s مجلة كلية التنمية البشرية
|t Journal of Faculty of Human Development
|v 000
|
856 |
|
|
|u 1432-000-004-010.pdf
|
930 |
|
|
|d y
|p y
|q n
|
995 |
|
|
|a IslamicInfo
|
995 |
|
|
|a HumanIndex
|
999 |
|
|
|c 1020426
|d 1020426
|