Semantic Question Answering of Umrah Pilgrims to Enable Self-Guided Education

978-1-4799-3516-1/13/$31.00 © 2013 141

Semantic Question Answering of Umrah Pilgrims to Enable Self-Guided Education

Nurfadhlina Mohd Sharef, Masrah Azrifah Murad, Aida Mustapha

Faculty of Computer Science and Information Technology, University Putra Malaysia,

Serdang, Malaysia {nurfadhlina, masrah, aida_m}@upm.edu.my

Saman Shishehchi Faculty of Technology and Information Science,

National University of Malaysia, Bandar Baru Bangi, Malaysia [email protected]

Abstract— Umrah as a pilgrimage to Mecca is obligatory in Islam madhab. Pilgrims can gain their knowledge about the requirements of Umrah using books, expert and existent question answering systems but still they suffer from the lack of complexity and natural language patterns. This paper proposes the semantic based question answering system to able pilgrims to compose any question about Umrah in natural language format. The proposed system used ontology to represent the knowledge about ritual of Umrah and Pilgrims. The question complexity for question in natural language is observed since it needs to be map with the contents in the ontology. An Umrah Knowledge module in this system covers all the rules and fact in the ontology format and the Educational modules is responsible for question answering interaction.

Keywords-Semantic question answering; Ontology, Umrah

I. INTRODUCTION Umrah is a pilgrimage to Mecca that may be performed at

any time of the year. Umrah pilgrims must be well prepared before going to Mecca. This involves knowledge such as on the requirements and conditions for each ritual of Umrah.

Questions may arise in challenging situations, and may happen to both first-timers and otherwise. Pilgrims may want to ask question pre, during or post-pilgrimage. They may want to know information about the traveling procedure, requirements and conditions of rituals and arrangements to return home. For example, “Considering that Hajj and 'Umrah are both obligatory in our madhab, if one has the choice of performing one of them, which should take priority?”. There are numerous online forums that provide collaborative support, however, the respond may be delayed or are sometimes unattended.

However, accessing information from experts may be constrained in terms of time, availability and cost. Although printed materials like books are available, the knowledge may not fit with the request of the pilgrim. Friends and accompanies may be around but they may be unsure of some information according to the level of their Umrah knowledge. The web can also be referred as a common resource of knowledge but the search engine does not have the capability to produce answer. Instead it typically returns a set of documents that contain occurrence of the searched keyword which may be overload to the user.

Works closely related are created specifically for Hajj which is the fifth pillar of Islam that is compulsory to able-bodied Muslims at least once in their lifetime. This may be the

main cause that has not motivated other researchers in exploring solutions for Umrah pilgrims. However, Umrah is regarded as the mini Hajj, and share some of the rituals of Hajj. For this reason we are motivated to explore on suitable approaches to assist Umrah pilgrims by utilizing semantic web (SW) technology. Furthermore, existing works on SW for Quranic related topics were focused mainly on solah information representation [1], [2], information retrieval [3] and computational linguistics [4].

Therefore in this paper we introduce a preliminary work on semantic-based question answering (SQA) system to enable self-guidance of Umrah education. The ontology based application delivers information by answering the question. This approach does not only provide accurate answer but also allows user to compose question in their natural language.We also address the issues relating to SQA for Umrah.

The paper is organized as follows. The first section of the paper introduces the work while the second section focuses on the existing works. The third section details the SQA for Umrah followed by the discussion in section four and conclusion in the final section.

II. EXISTING WORKS

A. ICT-based Solutions for Hajj Problems Some Umrah package providers provide trainings to

pilgrims before departing for Mecca in order to assist them. However, they probably encounter related problems that require them to make immediate decisions especially while performing the Umrah. Several works on ICT-based solutions for Hajj are available, which may have been motivated because of its importance to Muslim and because of the larger number of benefiters. However, we argue that ICT for Umrah is equally important to make sure that the pilgrims performed their best. Tto the best of our knowledge, there is so far no available information communication technology (ICT)-based applications developed to assist Umrah pilgrims.

Analytical study on potential ICT solutions to serve Hajj is provided in [5] where a Comprehensive Hajj Model is proposed. The model comprises of a multi-lingual Web-based Educational model for emergency and self-guidance. A database model for managing, organizing, and tracking pilgrims and other stakeholders is also included. Although the design of the model is quite extensive, it is not yet implemented therefore reliability cannot be guaranteed.

142 2013 13th International Conference on Intelligent Systems Design and Applications (ISDA)

Layered problems on Hajj rituals was studied in [6] where Hajj queries expert system is proposed. Knowledge based approach was created to improve pilgrim’s understanding of Hajj rituals. The rules allow question that involves condition checking or hukm. The prototype called Hajj-QAES is introduced in [7]. This application is targeted for the experts to capture both simple and advanced questions. This QA application is similar to our proposed work in the sense that facts and rules are encoded to solve questions. On the other hand, our proposed work utilizes ontology which allows the representation of conceptual knowledge and we focus specifically for Umrah.

A question answering (QA) system for Hajj Complex problems based on the hybrid of rule and frame approach is presented in [8]. Four variables are utilized which are question, answer, solution and user’s selected answer. Examples of complex problems are model in the frame and the rules are added to notify the sequence and requirements of each step in the rituals. The effectiveness of the model is compared against user experience on using the books, Internet or training modules to find the answer to similar problems. Results have shown that the model has reduced searching time and concluded that static websites do not cater the needs for complex and layered problems. However, this paper also does not address Umrah problems.

Another similarity shared by all the mentioned works is that there is none that approach the problem with semantic technology. Besides, there is no study done on the complexity and natural language patterns that naturally is used by the user to compose the question. The next section discusses related works on semantic QA.

B. Semantic Question Answering SQA is a discipline that allows the development of smart

systems through its reasoning and natural language understanding capability. Knowledge is stored based on representation format such as ontology. However, querying the knowledge requires understanding of the knowledge structure and demands the user to be equipped with the formal query (i.e., SPARQL) construction skill.

Therefore, casual user preferred an interface that use natural language (NL) compared to those keyword based, partial sentences and graphical [9]. However, processing NL questions are challenging due to the inherited ambiguity in the language. Question processing was approached based on several techniques; (i) interrogative question header [10], (ii) controlled natural language [11], (iii) focus and type of answer [12], and (iv) number of triples [13]; where important expressions that carry hints for the answer is detected and then translated into formal query format. Besides understanding the question, another challenge remains, which is to map the expressions in the questions with the ontology concepts.

For example, the question ‘Considering that Hajj and 'Umrahh are both obligatory in our madhab, if one has the choice of performing one of them, which should take priority?’ is very complex because it is constructed of several sub-sentences. This question can be rephrased according to several complexity level; (i) Should Hajj be prioritized against

Umrah? (ii) Which should be prioritized between Hajj and Umrah? (ii) If one has the choice of performing one of Hajj and Umrah, which one should be prioritized?

Automatically processing and translating this question [14] to the formal query requires several processes such as (i) stemming and stop word removal, (ii) part-of-speech tagging, (iii) question identification, (iv) sub questions identification, (v) question’s term to ontology concept annotation, (vi) formal query operator identification, (vii) formal query construction, (viii) query execution. Several NL processing tools will be needed to understand and translate the NL question to the formal query.

Besides the standard processes, the translation task is also hindered by the ambiguity in the question understanding and question’s term to ontology concept mapping. Existing works on the question understanding have focused mainly on the question focus extraction but have not explored much on the correct mapping of the terms in the question with the ontology concept. The main factor is because of the ambiguity in the mapping which is solved by the clarification dialogue approach where user is involved to select the correct concept. This is overloading the user and may require customization.

Therefore, in this paper we present the preliminary work of S QA for Umrah where the focus is given to the complexity types of the questions. We present the ontology design which is the heart of the automatic translation from the NL question to the formal query. Although our focus is specifically on Umrah, we share the same motivation by earlier works that encode Quranic knowledge in the form of semantic representation [1–3].

III. SEMANTIC QUESTION ANSWERING FOR UMRAH We extend the identified weakness and difficulties related

to Hajj [5] which is also shared by Umrah pilgrims which are the motivation for the work presented in this paper. They are the lack of comprehensive database for Umrah related stakeholders, lack of knowledge, misunderstanding, or wrong understanding of Umrah activities and rules by some of the pilgrims, agents, guiders and organization committees and lack of comprehensive and trusted education systems for Umrah.

Therefore, in this paper we address an ICT specification in the form of semantic QA for Umrah which can be the self-guidance tool for the pilgrims, agents, guiders and organization committees. This section describes the modules in the semantic QA and the complexities of the NL question for Umrah problems.

A. Architecture We propose a Semantic Question Answering (SQA) for

Umrah that consists of two modules, namely (i) Umrah Knowledge Module (UKM) where all facts and rules are stored in the form of ontology (see TABLE I) and (ii) Umrah Educational Module (UEM), which allows the question answering interaction to be executed.

Knowledge for this research is gathered from several resources such as books, the Internet and Umrah experts. Since the SQA is ontology-based, we follow the standard

1432013 13th International Conference on Intelligent Systems Design and Applications (ISDA)

ontology development methodology which is a cycle of identification of the following; competency questions, key concepts, important terms and terms relationship.

The facts in the UKM module are stored in the form of <subject,relation,object>. Each concept (subject/object) are connected through the relations. The module contains information such as definition, types, rituals, and distance.

TABLE I: Umrah Knowledge Module

Subject Relation Description Umrah hasDefinition Pilgrimage to Mecca Umrah performedBy Muslims Umrah performedWhen All year round Umrah hasRituals Tawaf, sa’ee, taqsir Umrah peakTimes Days before, during and after Hajj

and during last ten days of Ramadhan Umrah hasTypes Al mufradah, al-tammatu Ihraam hasMethod Perform ghusl, perform nafl, niyyah,

talbiyah ForbiddenAction validityIs Sexual intercourse, kiss, embrace,

unlawful act, hunting, scent, shave, clip hair, men use stitched clothes, men wear footwear that covers central bone of the upper part of the feet, men cover heads and faces, women cover faces, cause harm to any muslim

Tawaf hasMethod Proceed towards the corner of Kaabah, locate and kiss Hajarul aswad, make idhtibaa, make niyyah, keep Kaabah on your left and circle the Kaabah 7 times, istilaam, sa’ee, 2 rakaat prayer at Maqaami Ibraaheem, visit Zamzam well, visit Multazam

Tawaf hasDefinition Circling Kaabah 7 times in anticlockwise

Tawaf hasDistance Approx. 200m to 2 km Sa’ee hasMethod Climb safa, niyyah, raise hands, read

durood shareef, make du’aa, walk to marwah

Sa’ee hasDefinition Rapid walking back and forth safa and marwah

Safa hasDistance Approx. 3.5 km Halaq hasDefinition Cutting hair Halaq hasMethod Shave head or trim hair

The UEM module allows three main categories of questions as shown in TABLE II. Each question category allows users to pose question for different purpose as listed in the column labeled ‘Specification’.

TABLE II: Question Category

Question Category

Specification Examples

Definition Information What is Umrahh? Explanatory Person Who should perform Umrahh?

Time When should Umrahh be performed?

Information What are the rituals of Umrahh? Distance How far do we have to walk for

the tawaf? Steps How should tawaf be performed?

Affirmative Yes/No Is hair cutting prohibited during ihram? Can one wear perfume during ihram?

B. Ontology Design The main component in the SQA is the ontology. Before ontology is build, the knowledge source should be identified. This can be sourced from experts and reading materials. The scope of the SQA is determined by the coverage of the ontology. For example, the information extracted in Figure 1 is used to create the ontology as in Figure 2 .

Figure 1: Knowledge for Umrah1

Note the network of the concepts consists of six classes,

namely Umrah, Definition, DefinitionOwner, performTime, UmrahSequence and Ritual. The square boxes that are annotated by the diamond shapes are individuals, or also known as instances. The ‘Arc Types’ as shown in the Figure 3 indicate the name of the connections, or the properties that connect the concepts.

However, note that the ontology that we developed have wider coverage, as explained previously in Section III (A).

C. Question Complexity The following are the popular searched topics on Umrah

such as: i. Umrah during Hajj

ii. How to perform Umrah iii. Where and when to put Ihraam, iv. Hair cutting during Umrah v. Health problems prevent Umrah

These topics are also addressed in several variants and combined with other topic such as Hajj; for example ‘Hajj first or Umrah first?’, and ‘Umrah before Hajj’. Naturally, these questions relate to the same intention and thus the same answer. These questions are simple and straight forward, as in ‘Who can be a Mahram?‘. However, there are also long and complex questions such as:

vi. I am saving up to go for Umrah Insha Allah, but one problem is that I have a few debts outstanding. Would it

vii. be Islamically immoral to go Umrah without paying these debts off?

viii. I’ve never been Hajj/Umrah but I am working in Saudi.

ix. Can I visit Makkah to see how it all is and not perform Umrah because I will be performing Hajj in a couple of weeks anyway?

1 * from http://en.wikipedia.org/wiki/Umrahh on 22 July 2013, 5.37 PM

The Umrahh or (Arabic: عمرةة ) is a pilgrimage to Mecca, Saudi Arabia, performed by Muslims that can be undertaken at any time of the year. In Arabic, Umrahh means "to visit a populated place". In the Sharia, Umrahh means to perform Tawaf round the Kaaba and Sa'i between Al-Safa and Al-Marwah, after assuming Ihram (a sacred state), either from a Miqat like Zu 'l-Hulafa, Juhfa, Qarnu 'l-Manāzil, Yalamlam, Zāt-i-'Irq, Ibrahīm Mursīa.


Figure 2: Ontology according to the Knowledge Extracted in Figure 1

Figure 3: Arc types for the Umrah ontology

These questions are categorized as affirmative questions

where the answer can be a yes/no. However, they are in a complex structure compared to the straightforward affirmative question ‘If a woman’s hair is prone to falling out, will she still have to pay a dam?’ because the actual questions are accompanied by several statements that set the context of the information asked in the question.

In some cases the contextual information can be ignored but there are also cases where they need to be considered. For

example, in question that relates to an illness where the contextual information entails the state of the illness.

In the perspective of automatic translation of NL question to formal query, the terms in the question should be mapped and annotated with the ontology concepts. Annotating the questions with the ontology concept is challenging because the terms may not be a direct match with the concept names in the ontology. A few solutions are available such as involving the user to disambiguate the annotation [15–17] and using WordNet [17] to find synonyms between the term and the matching ontology concepts.

The annotated inputs as shown in TABLE III assist the construction of matching ontology triples which are then arranged to construct a SPARQL. Suitable modifiers are inserted to ensure correct result manipulations as show in the example of NL question to SPARQL translation in

TABLE IV. In the scenario of automated NL question to SPARQL translation, the translator of the SQA should have the ability to identify correct concepts and their connections.

TABLE III: Annotation of Question with Ontology Concepts

<HajjClass>Hajj</HajjClass> <seq>first</seq> or <UmrahClass>Umrah</ UmrahClass> <seq>first</seq>? <UmrahClass>Umrah</UmrahClass> <seq>before</seq><HajjClass>Hajj</ HajjClass >? Who can be a <MahramClass>Mahram<MahramClass>?

1452013 13th International Conference on Intelligent Systems Design and Applications (ISDA)

TABLE IV: Examples of Translation of NL Questions to SPARQL Query

Natural Language Question SPARQL Query What is the definition of Umrah? SELECT ?o WHERE {:Umrah

:hasDefinition ?o.} Where can Ihram be performed? SELECT ?o WHERE {?s:miqatAt

?o.} What are the rituals of Umrah? SELECT ?o2 WHERE {?s:hasRituals

?o1. ?o1 :ritualIs ?o2.} How many rituals need to be performed during the Umrah?

SELECT COUNT(?o2) as count1 WHERE {?s:hasRituals ?o1. ?o1 :ritualIs ?o2.}

What is the first ritual of Umrah? SELECT ?o2 WHERE {?s:sequenceFirst ?o1. ?o1 :ritualIs ?o2.}

If I wear a niqaab is my Umrah still valid?

ASK WHERE {:wear_niqaab rdf:type :ForbiddenAction. }

What are the forbidden actions during performing Umrah?

SELECT ?s WHERE {?s :validityIs :Invalid }

IV. DISCUSSION SQA for Umrah poses potential in facilitating education

platform to the pilgrims. Even though the pilgrims made some preparation pre-arrival to ensure that they will perform the ibadah at their best, they may face some situations that require extra help such as to validate whether their action is permitted or not during Umrah. Furthermore, the situation may be spontaneous and leave the pilgrims with lots of uncertainty. The experts may be referred but not necessarily easily accessible.

A few issues are related to the SQA for Umrah. Firstly, we acknowledge that the depth of the knowledge can vary according to the referred source2. For example, alternative source can be obtained from a different source which may influence the design of the ontology and indirectly the scope of the SQA. Nevertheless, the semantic-based representation in the ontology format allows for deep knowledge expression which can be manipulated in the query process. The embedded reasoning engine enables inference to be performed thus mimicking the experts’ method to solve a question.

Secondly, the automatic translation of question to SPARQL is challenged according to the depth of the ontology. Besides this, ontology authors may use different terminology to represent similar things, and this may create ambiguity problem. In this aspect several studies in SQA are working on improving the ambiguity problem but the paper has addressed the minimal requirement towards development of an appealing SQA for Umrah.

The third issue is the scope of the SQA. Existing technology has no ability to process complex question, such as question on hukm. This is because, typically answering this type of question involves logic and some rules. Existing semantic web technology allows rules and constraints to be added, but some questions can be very tangled and require the federation of a few queries.

Automating this process may be very challenging although not impossible with enough training and modeling. This also may indicate the deep involvement of experts to help the knowledge engineer to model the Umrah knowledge, validate

2 Such as from http://www.hajtours.net/Umrah.pdf

the model and to check the correctness of the translated queries.

V. CONCLUSION In this paper we addressed the issues of SQA for Umrah pilgrims education. We introduced the ontology for Umrah and discussed the issues pertaining to the development. The SQA approach for Umrah has not been explored before and existing studies are limited to the solah concept only while other related works focused on information retrieval and computational linguistics aspects. The proposed Umrah SQA consists of the knowledge and education module. The issues that we have addressed are the scope of the ontology and SQA and the question complexity. For future work we will explore more variation of question complexity and enrich the ontology content.

ACKNOWLEDGEMENT

The authors gratefully acknowledge the funding support received from the Malaysian Ministry of Higher Education under the Long Research Grant Scheme, Project No: LRGS/TD/2011/UITM/ICT/01/02.

REFERENCES

[1] S. Saad, N. Salim, H. Zainal, and S. A. M. Noah, “A

framework for Islamic knowledge via ontology representation,” 2010 International Conference on Information Retrieval & Knowledge Management (CAMP), pp. 310–314, Mar. 2010.

[2] S. Saad, N. Salim, and H. Zainal, “Islamic Knowledge Ontology Creation,” 2009.

[3] A. Azman and M. A. Azmi Murad, “Quranic-based concepts: Verse relations extraction using Manchester OWL syntax,” 2012 International Conference on Information Retrieval & Knowledge Management, pp. 317–321, Mar. 2012.

[4] E. Atwell, C. Brierley, K. Dukes, M. Sawalha, and A. Sharaf, “An Artificial Intelligence approach to Arabic and Islamic content on the internet,” in National Information Technology Symposium (NITS), 2011, pp. 1–13.

[5] S. a. Hameed, “ICT to serve Hajj: Analytical study,” International Conference on Computer and Communication Engineering (ICCCE’10), no. May, pp. 1–7, May 2010.

[6] S. Sulaiman, H. Mohamed, A. Ahmad, and S. Sulaiman, “A knowledge-based approach to facilitate queries by Hajj pilgrims,” 2008 International Symposium on Information Technology, pp. 1–9, 2008.

[7] S. Sulaiman, H. Mohamed, M. R. M. Arshad, N. A. A. Rashid, and U. K. Yusof, “Hajj-QAES: A Knowledge-Based Expert System to Support Hajj Pilgrims in Decision Making,” 2009 International Conference on Computer Technology and Development, pp. 442–446, 2009.

[8] H. Mohamed, S. Sulaiman, and M. Sabudin, “A Hybrid of Rule and Frame Based Approach in Solving Hajj Complex Problems,” 2009 International Conference of Soft Computing and Pattern Recognition, pp. 466–471, 2009.

[9] E. Kaufmann and A. Bernstein, “How useful are natural language interfaces to the semantic web for casual end-users?,” in The Semantic Web, 6th International Semantic


Web Conference, 2nd Asian Semantic Web Conference, 2007, pp. 281–294.

[10] E. Kaufmann, A. Bernstein, and R. Zumstein, “Querix: A Natural Language Interface to Query Ontologies Based on Clarification Dialogs,” in Proceedings of the 5th International Semantic Web Conference (ISWC 2006), 2006, no. November, pp. 980–981.

[11] A. Bernstein, E. Kaufmann, and R. Zumstein, “GINO - A Guided Input Natural Language Ontology Editor,” in Proceedings of the 5th international conference on The Semantic Web, 2006, no. November, pp. 144–15.

[12] D. Damljanovic, M. Agatonovic, and H. Cunningham, “FREyA: an Interactive Way of Querying Linked Data Using Natural Language,” in Proceedings of the 8th international conference on The Semantic Web, 2011, pp. 125–138.

[13] M. Yahya, K. Berberich, S. Elbassuoni, M. Ramanath, V. Tresp, and G. Weikum, “Deep answers for naturally asked questions on the web of data,” Proceedings of the 21st International Conference Companion on World Wide Web - WWW ’12 Companion, pp. 445–448, 2012.

[14] N. M. Sharef, S. A. M. Noah, and M. A. A. Murad, “Issues and Challenges in Semantic Question Answering through Natural Language Interface,” Journal of Next Genaration Information Technology, vol. (accepted), 2013.

[15] D. Damljanovic, M. Agatonovic, and H. Cunningham, “Natural Language Interfaces to Ontologies  : Combining Syntactic Analysis and Ontology-based Lookup through the User Interaction,” in Proceedings of the 7th Extended Semantic Web Conference (ESWC 2010), 2010, pp. 1–15.

[16] M. Yahya, K. Berberich, and S. Elbassuoni, “Natural Language Questions for the Web of Data,” in Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, 2012, no. July, pp. 379–390.

[17] V. Lopez, M. Fernández, E. Motta, and N. Stieler, “PowerAqua: Supporting Users in Querying and Exploring the Semantic Web,” Semantic Web, vol. 3, no. 3, pp. 249–265, 2012.

Semantic Question Answering of Umrah Pilgrims to Enable Self-Guided Education

Documents