Lie detection through voice can be detected using
the algorithm bidirectional associative memory. This system
is a branch of sound processing that can be used to identify
the type of sound lies use some verbs like go, roads and move.
This study uses an algorithm bidirectional associative
memory for the process and the introduction of lie detection
training through the sound use of bidirectional associative
memory. The system was tested by simulating the training
data and test data to generate a percentage of voice
recognition and classification of these lies. Experiments
performed with several changes in parameter values to
obtain the best percentage of recognition and classification.
The highest level of recognition contained in the verb "go"
with up to 90%. Results of this research is a sound that
indicated not indicated lies and deceit in the form of values
are classified according to the type of sound that is known
from the results of calculations of energy use bidirectional
associative memory.
Published In : IJCAT Journal Volume 2, Issue 8
Date of Publication : August 2015
Pages : 301 - 306
Figures :04
Tables : 03
Publication Link :Lie Detection System with Voice Using
Bidirectional Associative Memory Algorithm
Bustami : Completed Bachelor’s Degree
in MIPA Mathematic at University of
Syiahkuala (UNSYIAH) and He
Completed a Postgraduate in Department
of Informatics at STMIK Eresha Jakarta.
He has written many books that have
been published throughout Indonesia.
Fadlisyah : Completed Bachelor’s Degree
in Computer Science at University of
Padjadjaran Bandung and He Completed
a Postgraduate in Computer Systems
Engineering at University North Sumatera
(USU). He has written many books that
have been published throughout
Indonesia. Who occupied positions
ranging Informatics Laboratories, Head of
Community Service, Head of Robot
Intelligence Study Center, and also as a
coach MTQ branch DAK Qur'an.
Cultivated fields of research interest is the Image Processing and
Speech Processing.
Bidirectional Associative Memory
Lie
Voice
The conclusions of this study are as follows:
a. Bidirectional associative memory can be used for lie
detection through voice.
b. Tests conducted on samples outside the training
resulted in the recognition rate of 90%.
c. The highest level of recognition contained in the verb
"go" with up to 90%.
[1] Fadlisyah, Bustami, M.Ikwanus. Voice processing. First
Edition. Yogyakarta. Graha Science Publishers, 2013.
[2] Kurniawan, Harry. Comparison Fast Fouier Transform
DenganDisrcrete Fourier Transform at voice, University
of Malikussaleh, 2013.
[3] R.H.Sianipar, I.K. Wiryajati, M.Irwan. Digital Signal
Processing. Yogyakarta.Penerbit Andi, 2012.
[4] Sariadi, Gender identification through Saara with
Discrete Fourier Transform Methods (DFT), University
of Malikussaleh, 2013. [5] Syahrial, Dahlan Abdullah, Fadlisyah. Classification of
gun through sound using wavelet transform In
detachment b pioneer village jeuleukat. University of
Malikussaleh, 2014.
[6] Valluru B. Rao.C++ Neural Networks and Fuzzy Logic.
MTBooks, IDG Books Worldwide, Inc. 1995.
[7] Matthew L. Newman, James W, Diane S. Berry,
Pennebaker, Jane M. Richards. Predicting Deception
from Linguistic Styles, The University of Texas at
Austin, Southern Methodist University, The University of
Washington.
