PENGENALAN HURUF KATAKANA MENGGUNAKAN METODE ELMAN RECURRENT NEURAL NETWORK
Daftar Isi:
- Katakana Japanese characters has its own difficulty level in writing for non-Japanese people. Katakana is very different from the commonly used letters of the alphabet, katakana is also difficult to remember by non-Japanese nation because of its unique shape. Therefore we need software that can help recognize katakana. Data taken as many as 20 people at each of the 46 letters and is written as much as 5 times of students Japanese language courses at Language Institute Magenta. The software is built using Elman Recurrent Neural Network as recognition methods and Haar Wavelet method as feature extraction method. The results of the study of handwriting recognition katakana produce an accuracy of 74.35%.