NEURAL FUZZY PETRI NET BASED ARABIC PHONEME CLASSIFIER WITH MFCC FEATURE EXTRACTION
##plugins.themes.bootstrap3.article.main##
پوختە
In this paper Arabic phoneme classification is employed using Mel Frequency Cepstral Coefficient (MFCC) as the basic recognition features. These features are first calculated, then used as an input to fuzzy neural Petri net. One network are used for each phoneme. The network was first trained based on a set of training recoded data, then the network are validated based on another set of data. Classification accuracy were then calculated and it have been found that the resulting total accuracy reached 74.94%.
##plugins.generic.usageStats.downloads##
##plugins.themes.bootstrap3.article.details##
چۆنییەتی بەکارهێنانی سەرچاوە
Qalaai Zanist Journal allows the author to retain the copyright in their articles. Articles are instead made available under a Creative Commons license to allow others to freely access, copy and use research provided the author is correctly attributed.
Creative Commons is a licensing scheme that allows authors to license their work so that others may re-use it without having to contact them for permission
