Out of vocabulary detection in Indonesian speech recognition using word and syllable level decoding

One of the problems in speech recognition is out of vocabulary words (OOV) because they can make some words error. Out of vocabulary words are the words that cannot be recognized by speech recognizer because there is no recognizing database. Alignment, language model, and POS Tag method is proposed in order to recognize word error because of OOV words. Word and syllable level decoding from speech recognizer is the input for this method. Alignment is applied to word and syllable level decoding to get some differences from word and syllable level decoding. After that, language model and tag are also applied to determine if the words are correct. Speech recognition accuracy is about 75% if OOV rate is 15,5%. The OOV detection process reaches about 87% precision and 75% recall. Experiments also show that by using OOV detection, speech recognizer accuracy is increased by 11%.