The Study of Inference Rule Pattern of Al-Quran Malay Text

Translation

Abstract: Backward and forward searching are rule based inference techniques use in providing a mean of
knowledge acquisition in Artificial Intelligence field. Both techniques will lead into the goal or the conclusion
which is called the knowledge. Between these two techniques the differences lies in the propagation of the
chain whether it is exhaustive search or not. When exhaustive search pattern is used time taken to reach to the
knowledge is longer as compared to the non-exhaustive search. This study is to investigate the pattern of these
searching techniques found in Sheikh Abdullah Basmeih Malay Text Translation of Al-Quran. The texts are
classified into backward and forward searching and were tabulated into the criteria laid out for each verses.
Then the finding is obtained and the result will be further refined in order to develop an expert system tool to
assist the propagation of knowledge from verses to another verses based on the pattern observed. This
hopefully will be a great help to better understand the Al-Quran as the book of knowledge.

Keywords: Inference technique, backward searching, Al-Quran, knowledge propagation, text pattern.
Introduction

Inference rule is technique makes use of production rule in a form of condition-action pair which can be
represented as a set of if..then statement, if (condition) then
….action
or
action then
if (condition).

The rules can be evaluated by forward chaining or backward chaining which are very important in
Artificial Intelligent research1. The importance of these techniques lies in its capability to provide a mean
for knowledge acquisition in a large knowledge base system. These rules are used by inference engine
which is later used to generate new facts that in return will enrich the knowledge gathered 2.

The Holy Quran is the reference book for more than 1.6 billion of Muslims all around the world. Extracting
information and knowledge from the Holy Quran is of high benefit for both specialized people in Islamic
studies as well as nonespecialized people. The Holy Quran is the word of God and hence needs careful
handling when processed by automated methods of machine learning, natural language processing and
artificial intelligence.

With the help of Artificial Intelligence techniques and approaches, there stands an opportunity to further
analyze the content of Al-Quran3. The analysis can take the view from various ways such as to study the
pattern by which the verses are arranged, the underlying knowledge of the verses, or the organization of
the whole chapters. In this study, the focus will be on Surah Al-Baqarah and the goal of the study is to
investigate the pattern of searching techniques found in Sheikh Abdullah Basmeih Malay Text Translation
of Al-Quran. Usually the forward chaining which is also known as exhaustive search will propagate
further as compared to backward chaining. Later the pattern observed will be used to produce a tool for
categorizing the verses. And also to prove the forward searching pattern propagates along the affected
verses in various chapters.

Literature Review

Studies on the use of inference techniques had long been conducted and still there is much more to explore
in this area. This is because the inference technique is the most basic technique in cognitive area. One of
the use of rule based method is to classify molecules on any chemical substances4 which resulted in
producing an ontology-mediated application to asisst people working in biomedical and life sciences field.
Another study attempted on producing the application of inference technique in the field of providing a
suitable punishment based on Islamic Syariah Law found in the Al Quran5. On top of that the use of expert
system had widen its scope from MYCIN which run on mainframe computer system giving advice on the
medical diognosis to a system that is available on a Cloud based system6 which giving advice on the
classification of hadith. The application of the basic inference technique evolved from one application to
another application spreding on many platforms. For example, the Hadith found on the WWW is clasified
and then applying the data mining onto it in order to provide wider exploration on the desired data7.

Al-Quran and Hadith are two most important resources of knowledge that become a reference to Muslims.
But as far as this study is conducted not many research done on the Al Quran as compared to the Hadith.
In a Malay Text Translation of Al-Quran, there are a lot of verses containing words in a form of Jika..Maka
or Maka ..Jika. This two pair of words are said to be a rule based system which can be further classified
as forward chainig or backward chaining respectively.

Methodology

The study is done on the Sheikh Abdullah Basmeih Malay Text Translation of Al-Quran since this is the
most reliable source of Malay Translation available. The focus is on the longest chapter which is Al-
Baqarah since this chapter is regarded as one of the most important chapter.

The text is copied into Notepad for the purpose of extracting the verses containing the rule based pattern.
For this purpose the technique called feature vector or bag of words is used where all the punctuation is
removed and all texts are converted to small letter. Each verse is separated by a token. The following
diagram summarizes the process of extracting the text from the Malay Translation of Al Quran:

Document Tokenization Stemming

Bag of words Feature Selection

Then the bag of words is feed into Java text reader to separate and producing two different texts; one file
is for forward chaining verses and another for the backward chaining verses.

Textfile1

Al Baqarah Java Text Reader

Textfile Textfile2
Generally the content of Al-Quran is cllasified into 12 categories such as:
● Iman
● Ilmu
● Bangsa-bangsa terdahulu
● Sejarah
● Al-Quran
● Akhlaq dan Adab
● Ibadah
● Makanan dan Minuman
● Pakaian dan Perhiasan
● Hukum
● Muamalat
● Jihad

The verses from the backward and forward searching pattern are matched against the above categories.
The lookup table used in this case is the indexing table for identifying Al-Quran verses category. The
matching pattern with the category of the verses is as follows:

Result and Discussion

It is found out from 25 verses involved in the classifcation only 10 verses which are equivalent to 40%
fall into backward searching pattern and the other 60% of the verses fall into forward searching pattern.
Interesting fact illustraing from the findings is the occurrence of forward searching pattern is more than
backward searching pattern. It means the searching will propagate exhaustively and the knowledge or goal
of the verses lies in same chapter but different verses or it may be in the other chapters.

35
30
25
20
15
10 Forward
5 Searching
0 Backward
Pakaian dan…
Bangsa-bangsa…

Makanan dan…
Akhlaq dan Adab

Muamalat
Sejarah

Hukum

Jihad
Ibadah
Ilmu

Al-Quran
iman

Searching

Further the propagation of verses for the forward searching will be studied to understand the criteria of
verses involved. Then the clustering of the verses will be developed and will be used to develop an expert
system tool to assist the propagation from verses to another verses based on the pattern observed in order
to better understand the Al-Quran.
 Conclusion

Since most of the forward searching text pattern mainly fall into the category of Iman, therefore it is to be
expected the classification will reside in the next various chapters. Hence, this indicates Iman is a vital
requirement of a Muslim whereby it should not be taken lightly as other baranches of Islamic knowledge.
Therefore, this area of studies deserved further works using more robust techniques in order to come up
with a better application that enable Muslim to understand and treat the Al-Quran with great respect as a
book of knowledge.

Reference

1. Al-Ajlan A. The Comparison between Forward and Backward Chaining. Int J Mach Learn Comput.
2015;5(2):106-113. doi:10.7763/IJMLC.2015.V5.492.
2. Haval B, Mzori S. Forward and Backward Chaining Techniques of Reasoning in Rule-Based Systems.
2015;(July).
3. Aldhlan KA, Zeki AMAM, Zeki AMAM. Datamining and Islamic Knowledge Exctraction : Alhadith As
a Knowledg. Technol Commun. 2010:21-25.
4. Magka D, Krötzsch M, Horrocks I. A rule-based ontological framework for the classification of
molecules. J Biomed Semantics. 2014;5(1):17. doi:10.1186/2041-1480-5-17.
5. Nada YA, Aljahdali S. Expert System for Islamic Punishment (ESIP). 2011.
http://repository.nauss.edu.sa/handle/123456789/55433.
6. Bilal K, Mohsin S. Muhadith: A cloud based distributed expert system for classification of Ahadith. Proc
- 10th Int Conf Front Inf Technol FIT 2012. 2012:73-78. doi:10.1109/FIT.2012.22.
7. Saloot MA, Idris N, Mahmud R, Ja’afar S, Thorleuchter D, Gani A. Hadith data mining and classification:
a comparative analysis. Artif Intell Rev. 2016;46(1):113-128. doi:10.1007/s10462-016-9458-x.