See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/311790860

Prototype expert system using bayesian network for diagnose social illness

Article  in  Journal of Theoretical and Applied Information Technology · November 2016

CITATIONS READS
2 190

4 authors, including:

Rahmad Kurniawan R. Yendra

State Islamic University of Sultan Syarif Kasim Riau State Islamic University of Sultan Syarif Kasim Riau
16 PUBLICATIONS   28 CITATIONS    50 PUBLICATIONS   47 CITATIONS   

SEE PROFILE SEE PROFILE

Ahmad Fudholi
Universiti Kebangsaan Malaysia
155 PUBLICATIONS   2,048 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Solar Dryer Systems View project

The Extreme Currency Exchange Rate in South East Asia Using Classical and Bayesian Methods View project

All content following this page was uploaded by Ahmad Fudholi on 17 May 2017.

The user has requested enhancement of the downloaded file.

 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

PROTOTYPE EXPERT SYSTEM USING BAYESIAN

NETWORK FOR DIAGNOSE SOCIAL ILLNESS
1
RAHMAD KURNIAWAN, 2AFRIZAL M. NUR, 3RADO YENDRA, 4AHMAD FUDHOLI
1
Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri
Sultan Syarif Kasim (UIN Suska) 28293, Pekanbaru, Riau, INDONESIA
2
Department of Quran and Hadith, Universitas Islam Negeri Sultan Syarif Kasim (UIN Suska) 28293,
Pekanbaru, Riau, INDONESIA
3
Department of Mathematics, Faculty of Science and Technology, Universitas Islam Negeri Sultan Syarif
Kasim (UIN Suska) 28293, Pekanbaru, Riau, INDONESIA
4
Solar Energy Research Institute, Universiti Kebangsaan Malaysia,
43600 Bangi Selangor, MALAYSIA
1
E-mail: [email protected], 2 [email protected], 3
[email protected], 4 [email protected]

ABSTRACT

The social illness has become main problem in Indonesia such as drug addiction, prostitution, game
addiction, and so on. Based on Indonesian government, through its anti-drug agency, the National Narcotics
Board (BNN) in 2015 the number of drug users are 5 million people. Drugs have caused for as many as
15,000 deaths each year in Indonesia. Peoples did not get information about social ills such as the types of
drugs that have the same characteristics. As a largest Muslim country in the world may be reduce the social
ills with the Islam approach. Recently, the integration of Islamic science and technology has become
popular. At times the people get obstacles to get information and advice directly from expert and scholars in
Islamic science. This paper is to develop a prototype expert system by providing prevention and best
treatment based on Quran and Hadith. We use some of the data for knowledge base obtained from Selat
Panjang, Riau. Based on the preliminary results, it has similar diagnoses with expert. A prototype expert
system using Bayesian Network has been successful and capability for early diagnosing and educating the
peoples in social ills cases based on Quran, Hadith and advice of experts but still has plenty room for
improvement.
Keywords: Bayesian Network; Expert System; Social Illness

1. INTRODUCTION discrimination, family disintegration, poverty,

homelessness and so on.
A social illness is a term synonymous with
Indonesia as a country with the largest
the phrase social conflict, social problem or social
Muslim population in the world may be reduce
issue. A social illness has become main problems
social ills with the Islam approach. Many of the
exists areas out of capital city when a condition is
scholars in Islamic science who could give
undesirable to some members of a community. Like
guidance, intelligence and advice to the younger
a diseases infection, the social illness also spread
generation to solve social ills. At times the people
quickly. In a society where the permissiveness and
get obstacles to get information and advice directly
promiscuity are encouraged to proliferate. The
from scholars. This study obtained the data from
pressure on individuals is very high indeed. The
areas out of capital city namely Kepulauan Meranti
youth generally lack maturity and are vulnerable to
Regency Selat Panjang. Selat Panjang is part of the
the social ills if they are overwhelming. Some of
province of Riau in Indonesia. Selat Panjang is
social ills include drug addiction, free sex, bullying,
separated by a narrow channel from Sumatra, and
crime, online games addiction racism, delinquency,

338
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

west of Singapore across the Strait of Malacca. medical science. Based on the preliminary
Poverty stops people in Selat Panjang to get advice experimental results, it can be concluded that the
from the scholars regularly. Thus, many people did implementation of Bayesian Network shows a
not get the necessary advice from an expert and promising result.
therefore they do not take the big problems in social
Nowadays, there are many expert systems
ills seriously. Online self-diagnosing is one of the
being used in Islam science such as Classification
solutions as the Internet is relatively cheap to get
of Ahadith, Using Social Network on Publishing
more information and cybercafé is one of the
and Serving Islam, expert system for Inheritance or
prevalent ways for poor people to search about their
Fara’id [1]. The scholars give treatment based on
difficulty. Based on data obtained from the police
the al-Quran and hadith. A method in Quran to treat
and hospital in Selat Panjang, it has been found
the social ills namely Al-adabiyijtima’iy method
alcohol, methamphetamine, marijuana, game
used for knowledge base. In this study, we used the
addictions and prostitution cases. Knowledge base
Bayesian Network method for inference engine.
acquired from the scholars in Islamic science.
However, self-diagnosis is prone to misdiagnosis One of the advantages in using Bayesian
that may endanger the users health and social if Network in this project is that Bayesian Network
users make a wrong decision. considered more popular as network-based
Bayesian inference is consistent when handled with
Actually, there are several online articles,
uncertainty. Bayesian networks can facilitate
magazines and health books containing about how
learning about the causal relationship between the
to recognize the drug abuses based on physical
variables [2] Bayesian Network Easy to be
symptoms. There are many symptoms, both
converted into tools for decision support to assist
physical and behavioral, that indicate drug use.
the management of natural resources [3].
Each drug has its own unique indications, but there
are some general signs that a person is using drugs. In addition, graphic in Bayesian Networks
Building an expert system for the social ills are to clearly show the relationship between different
challenging because it is required to create a components of the system. Therefore, Bayesian
knowledge base that is appropriate with desires of Network comfortable for researchers from different
scholars in Islam, doctors, psychiatrists and backgrounds to understand the concept of Bayesian
psychologists. Distinction with another cases, the Network [4]. The use of Bayesian inference can
expert system for diagnosing social issues based on readily updated for new knowledge and Bayesian
different content of knowledge base data that networks can produce a good prediction accuracy
consist of symptoms and treatment instruction [2, 5]. Bayesian Network used to determine which
based on medical science and Islamic science. It has type of social illness has the probability to be
many expert system integrated with Islamic Science effecting a user. Bayesian Network calculate the
as we have explained in the related works. chances based on existing symptoms and the system
Nevertheless, an expert system for social ills that will later on be able to appropriate preventive,
integrate with Islamic Science and medical science medical solutions advice and treatment based on
is still not exist. Thus, the main goal of this project Quran and Hadits to the type of illness accordingly.
is not just to develop a prototype expert system but
for educating the peoples in social ills cases based 2. RELATED WORK
on Quran, Hadith and advice of experts so that Lately, integration of Islamic science and
people will get advice and treatment appropriately technology had become popular. Kashif Bilal and
and as early as possible. In this project, we have Sajjad Mohsin presents a novel approach for the
developed a prototype expert system for social classification of the religious scriptures, the Hadith
issue. A prototype expert system had used Bayesian [6]. A novel approach has been used to simplify
Network. It has proved its capability for educating and digitize the task of Hadith classification.
the peoples in social ills cases based on Quran, Service Oriented Architecture (SOA) is used for
Hadith and advices of experts. Based on the Cloud computing compatibility, and to solve the
preliminary experimental results, it can be communication problems faced by the legacy Web
concluded that the implementation of Bayesian based distributed expert systems. The social
Network shows a promising result. On the other network one of the best ways to serve and publish
hand, we have been providing a medium for Islam. Beside that, students determine that the
educating the peoples regarding the risks of social existence of a religious person is very important as
issues, the information to avoid and treat of social a resource of Islamic information. Therefore,
ills based on the Quran, the Hadith, psychology and Da’wah persons must take care about social

339
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

network and be an active member there because We have stated in the paper about the
they have a big effect on the member as the limitation of this research. A self-diagnosis is prone
respondent mention by 80% [7]. Furthermore, to misdiagnosis that may endanger the user, health
Houssen Himeda had discussed web based expert and social if user make a wrong decision. Thus, the
system for Islamic inheritance law for all Muslims, main goal of this project is not just to develop a
and identify the rules of wealth distribution as prototype expert system but for educating and
stated in the Quran through a knowledge giving the awareness to the society about social ills
acquisition process with an expert in Faraid [1]. cases based on Quran, Hadith and advice of experts
so that people will get advice and treatment
There are several studies in the field of appropriately and as early as possible. Currently,
artificial intelligence, especially expert system that we just concentrate in drug abuse diagnosis
has been done. Meigarani [8] using Bayesian features. Meanwhile for the games addiction, the
Network to diagnose the disease. Bayesian Network prostitution, the sexual deviation cases and so on
used to diagnose leukemia by getting to the positive will be done in further research, because of the
and negative decisions. Bayesian Network was used complexity of rules and need more the data. This
for data analysis and knowledge of experts, study is not the final result of expert system in
especially on uncertainty. This method is used as a social ills, for encouraging results we need the
solution to a problem that is not certain. This additional experimentation and much data for
method can be used to create a model system for knowledge base (symptoms, kinds of social ills,
transformation between expert knowledge to treatment). It also requires full development of
computer in difficult medical cases [9-11]. expert systems in the future with a more
Bayesian networks can be prepared with decision comprehensive in Islam science and Medical
support tools in uncertainty case [12, 13]. Science. There has plenty work needs to be done
before a full-size expert system is constructed with
Bayesian Network became very popular in
standard expert systems components, improvement
the last decade because it is used for a variety of
of knowledge base by more associating the
intelligent system applications such as machine
statistical science, medical science or drug
learning, text processing, natural language
rehabilitation centers with the Islamic laws and
processing, speech recognition, signal processing,
scholars in order to avoid a misdiagnosis.
bio-informatics, medical diagnosis, forecasting,
mobile networks, and other intelligent system The following are the research
applications. A Bayesian Network consists of a methodology of this study. The construct stages of
directed acyclic graph of ‘nodes’ and ‘links’ that this study are based on the popular evolutionary
conceptualise a system. Bayesian Network is prototyping. The stages are as follows:
Probabilistic Graphical Models (PGM) which
consists of probabilistic and graph. Probabilistic
3.1 Literature Review
theory related to the data. The relationships
The literature review is a select analysis of
between nodes are described by conditional
existing research which is relevant to expert system
probability distributions that capture the
and social ills topic, showing how it relates to this
dependences between variables. Bayes Formula
study. It describes and validates how this research
[14]:
investigation may assist answer some of the
| questions or gaps that related in this study. This
P A|B (1)
step include reading and searching information
Where: regarding Bayesian Network, social issues on
A = a particular state, conditional on the evidence books, journals, and other scientific resources such
provided. as web pages and articles related. The literature
P(A∣B) = posterior review is a select analysis of existing research
P(A) = prior probability of the hypothesis which is relevant to expert system and social ills
P(B∣A) = likelihood topic, showing how it relates to this study. It
Joint distribution formula: describes and validates how this research
P A |B P A P B|A (2) investigation may assist answer some of the
questions or gaps that related in this study. This
3. RESEARCH METHODOLOGY step include reading and searching information
regarding Bayesian Network, social issues on

340
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

books, journals, and other scientific resources such For this experiment, 10-case sample extraction was
as web pages and articles related. made. Each case was then presented to the system
to obtain the corresponding diagnosis. This
3.2 Feasibility Study diagnosis was later compared to the one provided
This is where the problems are defined, the by the expert to determine whether the system’s
objectives are stated, the resources, the methods, diagnosis is correct or not.
the experts, the costs and the time frame are clearly
identified. This step is the requirement analysis 4. ANALYSIS
usually carried out in the system development life
cycle.
Here are some of the process of the
analysis which has been carried out:
3.3 Knowledge Acquisition
Knowledge acquisition is the step in this 4.1 Analysis of Interface
study for extracting, organizing and structuring Inference used for the construct of this
knowledge from the human experts to computer. expert system, which is Bayesian Network in the
This is often the major obstacle in building an diagnosis process of social ills.
expert system. Knowledge elicitation is also
conducted and test cases were prepared in this time.
The type of knowledge gathered in this stage is Feasibility Knowledge
factual knowledge. Literature review
study acquisition

3.4 Prototype Development

Build the structure of a Bayesian Network,
In this step, we selected the appropriate Prior, Conditional Probability, Joint Bayesian Network
method to address the problem. Bayesian Network Probability and Posterior
used in this study is for calculating the possibility
of people affected by the social ills. Knowledge
base obtained from experts such as doctor in A prototype expert Peleminary Validation testing by
system Testing experts
hospital and scholars in Islamic science. The
treatment and solutions are presented based on
Quran and hadith using Al-adabiyijtima’iy method.
Pleminary results
Analysis carried out in order to the system worked Documentation
analysis
accordance with expected. First, we build a small
system containing few of the features and it will
evolves into better system in few cycles.
Figure 1: The Steps to Build Expert System
3.5 Testing Bayesian Network provided the diagnostic
The system testing and validation is based
results use the rule based as the learning. The
on the black box method and User Acceptance Test
process was performed using the Bayesian Network
(UAT). UAT is conducted by using several cases of
method and calculating the probability of each
social ills. The pleminary results of a prototype
symptom from users. Subsequently, the final
expert system will be assessed by an expert. Based
diagnosis based on the probabilistic final process.
on the analysis and advise, the development step
The overview of implementation of Bayesian
will go back to step Knowledge Acquisition to
Network in a prototype expert system process is
correct and expand the factual knowledge on the
shown in Figure. 1.
basis of the domain expert’s comments. The
purpose of this cyclic development is to improve 4.2 Implementation of Bayesian Network
the quality of knowledge base and how the Bayesian network used to determine the probability
inference mechanism. of a user in kinds of social ills based on symptoms
or characteristics. There are several steps to
3.6 Pleminary Experiment implement Bayesian network that is shown in
The special purpose of the pleminary Figure. 2.
experiment is therefore to study the effectiveness of
a cooperation Bayesian Network for a prototype
expert system for social ills. The system has been
assessed considering the accuracy as the indicator.

341
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

Build the Parameters/
Conditional
4.2.3 Conditional probability
structure of a Prior The conditional probability of an event B
Probability
Bayesian Network Probability is the probability that the event will occur given the
knowledge that an event A has already occurred.
Joint
Probabilistic Posterior This probability is written in formula (1) where
Probability
Inference Probability
Distribution event A has no effect on the probability of event B,
the conditional probability of event B given event
Figure 2: Analysis of Bayesian Network as Inference A is simply the probability of event B, that is P(B).
Engine
Table 2: Example of Conditional Probability Table
(CPT).
4.2.1 Build the structure of a bayesian Meth Social issue
network mouth
Present Absent
The structure of Bayesian network formed Present 0.7 0.25
using graph theory that connects the symptoms or
characteristics with the kind of social ills. Bayesian Absent 0.3 0.75
network structure as shown in Figure 3.
dating excessively for the
ability to have multiple Seclusion, nasty
4.2.4 Joint probability distribution
excessive
dressing
(tabarruj)
sexual partners
sexually harassing
(khalwat/ikhtilat)
with strangers (not We have used the formula (2) Joint
a mahram)
others/rape/
molestation distribution formula to obtain the Joint Distribution
bloody or black/
vomiting headache
chronic
compulsive masturbation
or stimulation Table (JPT). Based on the formula (2), the
diarrhea
clumsiness
extra-marital affairs or calculation of joint probability distribution is
Social Issues multiple affairs in a non-
gastritis
difficulty
walking
married relationship multiplying the prior probability Alt conditional
redness of the Alcohol hallucinations
Zina/
Prostitution probability. Suppose the joint probability
nose and
cheeks Khamar/
Drugs
distribution will be calculated in meth mouth. The
increase of
sickly sweet
smell on anxiety
“meth mouth” meth mouth present probability is 0.7, while absent
appetite body, hair extreme
and clothes
sleepiness
lethargy
blood pressure
0.3. The Joint Distribution Table (JPT) is shown in
uncontrolla Marijuana
will increase
Table 3.
ble dilated
laughter pupils
Crystal Meth Table 3: Joint Distribution Table (JPT)
weight extreme
red eyes gain or weight loss
loss of total loss
loss violence
interest of appetite
Meth Social issue
over
excited mouth
Game
Addictions
Present Absent
lying to
friends or
Present 0.3 × 0.7 = 0.21 0.7 × 0.25 = 0.175
family
spend anticipation of
more time
gaming
the next online
session
. Absent 0.3 × 0.3 = 0.09 0.7 × 0.75 = 0.525

Figure 3: The Structure of Bayesian Network with

Certain Symptoms and Kind of Social Ills Based on 4.2.5 Posterior probability
Expert Knowledge In statistical terms, the posterior
probability is the probability of event A occurring
4.2.2 Prior probability given that event B has occurred. Based on The Joint
Prior probabilities are the original Distribution Table (JPT) in meth mouth, The
probabilities of an outcome, which be will updated posterior probability is:
with new information to create conditional, joint
and posterior probabilities that is shown in Table 1. 0.21
0,54
0.21 0.175

Table 4: Example of Posterior Probability.

Table 1: Example of Prior Probability.
No. Symptoms Values
No. Symptoms Values
1. Total loss of appetite 0.54
1. Total loss of appetite 0.5
2. Meth mouth 0.4
2. Meth mouth 0.3
3. Dilated pupils 0.67
3. Dilated pupils 0.25

342
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

4.2.6 Probabilistic inference Table 5: Expert System results.

Cases Diagnosis by Expert System Diagnosis by
The Probabilistic inference is done by and Probability Human Expert
tracing the relationship of each symptom and kind 1 Marijuana (63%) Marijuana
of Social ills based on Bayesian Network Structure. 2 Marijuana (46%) Marijuana
3 Crystal Meth (55.5%) Crystal Meth
Sometimes, BN combined with Rule-Based
4 Game Addictions (52.5%) Game Addictions
Reasoning (IF-THEN) for assist the probabilistic 5 Crystal Meth (30%) Crystal Meth
inference [8]. Example: 6 Alcohol (41%) Alcohol
7 Crystal Meth (42%) Crystal Meth
8 Crystal Meth (45%) Crystal Meth
P (Crystal Meth |Symptoms of Crystal Meth)
9 Marijuana (77.5%) Marijuana
10 Marijuana (48.25%) Marijuana
0.54 0.4 0.67
0,53
3
Regarding the preliminary results that
Thus, its can be concluded that a people uses have used some of the knowledge base, it has
crystal meth that is with belief percentage is 0.53 × similar diagnoses with human expert. The diagnosis
100 % = 53 %. is more accurate if the user selects many of the
symptoms. Thus the prototype expert system results
according to the expected.
5. IMPLEMENTATION 7. CONCLUSION AND FUTURE WORK
The construction phase does two things:
constructs and assessments a functional system that We have developed a prototype expert
fulfills organizational design requirements, and system for social issue. A prototype expert system
implements the interface between the new expert had used Bayesian Network. It has proved its
system and the existing system. We have capability for educating the peoples in social ills
constructed the knowledge database, inference cases based on Quran, Hadits and advice of experts.
engine, application programs, user and system Based on the preliminary experimental results, it
interfaces, and networks. We developed the system can be concluded that the implementation of
based on the results of the previous model that has Bayesian Network shows a promising result.
been designed, so that it can be used in the real However, this study is not the final result of
situation. expert system in social ills, for encouraging results,
additional experimentation and much data for
6. EXPERT SYSTEM TESTING knowledge base (symptoms, kinds of social ills,
treatment) are needed. It also requires full
System testing and evaluation is the final development of expert systems with a more
phase in this study. System testing and evaluation is comprehensive in Islam science and holistic. There
ran to evaluate the effectiveness of the system. The has plenty work needs to be done before a full-size
develop system has to produce the desired outcome expert system is constructed with standard expert
in accordance with its objectives with high systems components, improvement of knowledge
accuracy without any systemic error. Testing are base by more associating the statistical science,
conducted using the black box method. medical science or drug rehabilitation centers with
User acceptance test was conducted to the Islamic laws and scholars in order to avoid a
validate the system output directly to the experts misdiagnosis.
and end-users. 10 cases have been selected and
assessed and tested by experts. The following table ACKNOWLEDGMENT
compares diagnostic results by expert systems and
the human expert. Authors would like to thank the scholars in Islamic
Science as experts, the police, hospital and social
services in Selat Panjang, Riau, Indonesia who has
guided this study and provided precious data on
social ills.

343
 Journal of Theoretical and Applied Information Technology
30th November 2016. Vol.93. No.2
© 2005 - 2016 JATIT & LLS. All rights reserved.

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

REFRENCES: [10] C. Papaconstantinou, G. Theocharous, and S.

Mahadevan, “An Expert System for Assigning
[1] H.H.M. Nafed, “Web Based Expert System Patients into Clinical Trials Based on Bayesian
For i-Faraid,” University Utara Malaysia, Networks,” Journal of medical systems, Vol.
2009. 22, 1998, pp. 189-202.
[2] L. Uusitalo, “Advantages and challenges of [11] P. J. Lucas, N. C. de Bruijn, K, Schurink, and
Bayesian Networks in Environmental A. Hoepelman, “A Probabilistic and Decision-
Modelling”, Journal of Ecological Theoretic Approach to the Management of
Modelling, Vol. 203, 2007, pp. 312-318. Infectious Disease at the ICU”, Artificial
[3] B.G. Marcot, R.S. Holthausen, M.G. Raphael, Intelligence in Medicine, Vol. 19, 2000, pp.
M.M. Rowland and M.J. Wisdom, “Using 251-279.
Bayesian Belief Networks To Evaluate Fish [12] S. Kuikka, M. Hildén, H. Gislason, S.
And Wildlife Population Viability Under Land Hansson, H. Sparholt, and O. Varis, “Modeling
Management Alternatives From An Environmentally Driven Uncertainties in Baltic
Environmental Impact Statement”, Journal of Cod (Gadus morhua) Management by Bayesian
Forest Ecology and Management, 2001, pp. Influence Diagrams”, Canadian Journal of
153: 29-42. Fisheries and Aquatic Sciences, Vol. 56, 1999,
[4] F.J. Martín De Santa Olalla, A. Domínguez, A. pp. 629-641.
Artigao, C. Fabeiro and J. F. Ortega, [13] T. D. Nielsen, and F. V. JENSEN: “Bayesian
“Integrated Water Resources Management of Networks and Decision Graphs. Springer,”
The Hydrogeological Unit “Eastern Mancha” 2009.
Using Bayesian Belief Networks”, Journal of [14] F. Markowetz, “Learning in Bayesian
Agricultural Water Management, Vol. 77, Networks,” Institute for Molecular Genetics,
2005, pp. 21-36. Computational Molecular Biology, Berlin,
[5] J.L. Ticehurst, R.A. Letcher and D. Rissik, June 2012.
“Integration Modelling and Decision Support:
A Case Study of the Coastal Lake Assessment
And Management (Clam) Tool”, Journal of
Mathematics and Computers in Simulation,
Vol. 78, 2008, pp. 435-449.
[6] K. Bilal, and S. Mohsin, “Muhadith: A Cloud
Based Distributed Expert System for
Classification of Ahadith”, Frontiers of
Information Technology (FIT), 10th
International Conference on 2012, IEEE, 2012,
pp. 73-78.
[7] I. A. Nassar, J. A. Hayajneh, and M. K.
Almsafir, “The Influence of Using Social
Network on Publishing and Serving Islam: A
Case Study of Jordanian Students”, Advanced
Computer Science Applications and
Technologies (ACSAT), IEEE, 2012 pp. 502-
505.
[8] I. Meigarani,“Penggunaan Metode Bayesian
network Dalam Sistem Pakar Untuk Diagnosa
Penyakit Leukimia”, Jurnal Universitas
Pendidikan Indonesia. 2010.
[9] J. C. E, Kahn, L. M. Roberts, K. A. Shaffer,
and P. Haddawy, “Construction of a Bayesian
Network for Mammographic Diagnosis of
Breast Cancer”, Computers Biol. Med. Vol. 27,
1997, pp. 19–29.

344

View publication stats