Kurniawan, Ferdinandus Hanry (2016) PATTERN RECOGNITION WITH BACKPROPAGATION ALGORITHM. Other thesis, Fakultas Ilmu Komputer UNIKA Soegijapranata.

Preview	Text (COVER) 12.02.0043 Ferdinandus Hanry Kurniawan COVER.pdf Download (276kB) | Preview
	Text (BAB I) 12.02.0043 Ferdinandus Hanry Kurniawan BAB I.pdf Restricted to Registered users only Download (33kB)
	Text (BAB II) 12.02.0043 Ferdinandus Hanry Kurniawan BAB II.pdf Restricted to Registered users only Download (207kB)
	Text (BAB III) 12.02.0043 Ferdinandus Hanry Kurniawan BAB III.pdf Restricted to Registered users only Download (90kB)
	Text (BAB IV) 12.02.0043 Ferdinandus Hanry Kurniawan BAB IV.pdf Restricted to Registered users only Download (108kB)
	Text (BAB V) 12.02.0043 Ferdinandus Hanry Kurniawan BAB V.pdf Restricted to Registered users only Download (523kB)
	Text (BAB VI) 12.02.0043 Ferdinandus Hanry Kurniawan BAB VI.pdf Restricted to Registered users only Download (28kB)
Preview	Text (DAFTAR PUSTAKA) 12.02.0043 Ferdinandus Hanry Kurniawan DAFTAR PUSTAKA.pdf Download (29kB) | Preview

Abstract

Complex pattern recognition is a difficult problem to solve because in the real v&rid patterns / images to be recognized has noise. The unavailability of good mathematical calculation that can produce the desired translation, nauseating to be able to resolve a problem of pattern recognition. Imagine if you have to match in the database one by one that there is also troublesome, as resolved by adding a large number of programs and inefficient. Therefore to solve the problem, back propagation algorithm is the best answer to a complex pattern recognition. Backpropagation algorithm is implement at Pattern Recognition Application. Backpropagation v\orks far faster than earlier approaches to learning, making it possible to use neural nets to solve problems v\ftich had previously been insoluble. Today, the backpropagation algorithm is the workhorse of learning in neural networks. This algorithm do a tv\o-stage calculation. First forward propagation to calculate the error betmen the actual output and the target Second, backmrd propagation, vtiich propagates the error back to fix the v&ights on all existing neurons. By using the back propagation algorithm, pattern recognition problems more easily and achieved results better than using matching algorithm. Because the character recognition using the data containing noise and sizeable

Actions (login required)

View Item