@article{MAKHILLJEAS202015419025,
    title = {Advance Natural Arabic Communication Recognition Uses a
Speech Recognition Approach},
    journal = {Journal of Engineering and Applied Sciences},
    volume = {15},
    number = {4},
    pages = {913-924},
    year = {2020},
    issn = {1816-949x},
    doi = {jeasci.2020.913.924},
    url = {https://makhillpublications.co/view-article.php?issn=1816-949x&doi=jeasci.2020.913.924},
    author = {Mohammed Basil},
    keywords = {Arabic language,speech,recognition,natural,Mel Frequency Cepstral Coefficient (MFCC),Dynamic Time Warping (DTW),classification},
    abstract = {In the era of natural computing, automatic recognition of effective natural communication database
for recognition is having major challenges. Human computer interface is always proved to be stimulating area
in the speech and the language recognition domain. Speech is the important choice of man-machine
communication. The objective of the speech recognition is to cultivate statistical methods and its use in real
time automatic speech recognition systems. The statistical model of speech recognition plays an important role
in the development of real time system. Arabic is a global language for Iraq country and has universal
acceptance. This is an official language at many institutes in Iraq. This study addresses the implementation of
an Advance Speech Recognition system (ASR) for an Arabic language towards identification of natural
communication aspects. The statistical modeling and implementation of the system is done in this study. This
research contributed the development, Arabic speech database for isolated (character, digit, word) and
connected sentences which are used day to day activity in Arabic community. In this study, 10 subject databases
were developed using the PRAAT Software in an office environment. The size of collecting database is
consisting of 1200 samples of isolated words and 800 samples of connected sentences, the size of the database
is 2000. The experimental analysis of Arabic natural communication is tested using the Mel Frequency Cepstral
Coefficient (MFCC) feature extraction technique and dynamic time warping algorithm. The isolated natural
recognition system gives 94% accuracy and continuous spoken natural communication proves the 87%. The
Natural isolated recognition system gives a robust and dynamic performance than continuous speech
recognition. This natural communication based emotional spoken database will be used for the emotive aware
cloud computing and spoken interface domain.}
    }