You are on page 1of 6




Analyze Semantic of Object-Oriented Model Using RiTa.WordNet
Noraida Haji Ali, Noor Syakirah Ibrahim, Nor Fazlida Mohd Sani and Noor Maizura Mohamad Noor
Abstract— The importance of modeling is obvious with its use to predict cost and time requirement especially in completing a system. However, there is no equivalent standard for evaluating the quality of conceptual models. Thomasson has also shown the difficulties in designing the appropriate UML class diagram such as naming the notation element. The UML class diagram designed by students always neglects the quality in modeling such as consistency and accuracy. This study proposes the use of WordNet in order to achieve the quality in modeling. RiTa.WordNet will be used as a tool to extract synsets from WordNet. The use of RiTa.WordNet shows that synonyms extracted can be used to match the UML class name designed by students. This application will be used to increase the accuracy of object-oriented model. Index Terms—Framework, Object-Oriented Model, RiTa.WordNet, WordNet

——————————  ——————————



RAMEWORK is a basic conceptual structure that used to solve or address complex issues, usually a set of tools, materials or components. It is also a reusable and halfcomplete application which used for producing other application [1], [2]. In a software context, the framework is used as a name for a different kind of tool sets. Currently, frameworks are most commonly represented through design diagrams written in standard object-oriented analysis and design languages. Frameworks used to model a specific domain or an important aspect of that domain [3]. They represent the domain as an abstract design, consisting of abstract classes (or interfaces). The abstract design is more than a set of classes, because it defines how instances of the classes are allowed to collaborate with each other at runtime. Effectively, it acts as a skeleton, or material that determines how framework objects relate to each other. Towards the system completion, modeling is essential to predict cost and time requirement. However, there is no equivalent standard for evaluating the quality of conceptual models. The traditional focus of software quality shows that only final product has been evaluating. The main task in objectoriented is concentrated on the construction of a model of a problem domain, rather than software implementation. Improving the quality of conceptual models is as important as to improve the quality of delivered system [4]. Traditionally, system modeling can be represented by text or graphically. Nevertheless, statement through this approach can solve some problems related to understanding the system requirements. This scenario will emerge the misunderstanding between user and system engineer. This will also make a system inconsistent [5]. Because of its significant popularity and is the de facto for

modeling software architecture and design, Unified Modeling Language (UML) was adopted as a standard by the Object Management Group (OMG) in November 1997 and now serves as the standard language of blueprints for software [6]. In modeling, designing the UML class diagram is an important phase. Nevertheless, the formal semantics in UML is still lacking, i.e. the meaning of the elements of a UML model is not formally defined and may depend on the interpretation of individuals who are using the UML [7]. Evaluating some problems in modeling, research from Thomasson shows the difficulties in designing the appropriate UML class diagram [8]. They are 1. The variation of the design form. 2. Naming the notation element. 3. Free in designing. 4. Difficult to state the class or object. 5. Difficult to elaborate the requirement. The UML class diagrams designed by students always neglect the quality in modeling such as consistency and accuracy. This should be overcome to make sure that there is no duplicate in class naming and the inheritance relationship is valid. For this purpose, a framework will be developed to overcome the inconsistency problem in UML class diagram [9]. The remainder of the paper is organized as follows. Section 2 discusses about the background of this research. The next section describes the method that is used in this study. Section 4 discusses the framework and presents result testing in Section 5. In section 6, we give the results of our study and the discussion in section 7. Section 8 summarizes the study and points out some future research issues.

Many ways have been done to produce a good system or model. One of that is by using tools such as CONCEIVER++, an understanding-based program debugger for object-oriented programming language [10]. However, analysis is also an important stage because the conceptual model can be shown to fulfill the requirements and becomes the skeleton to build a

 Noraida Haji Ali. Computer Science Department, Universiti Malaysia Terengganu, 21030 Kuala Terengganu, Malaysia.  Noor Syakirah Ibrahim. Computer Science Department, Universiti Malaysia Terengganu, 21030 Kuala Terengganu, Malaysia.  Nor Fazlida Mohd Sani. Computer Science Department, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia.  Noor Maizura Mohamad Noor. Computer Science Department, Universiti Malaysia Terengganu, 21030 Kuala Terengganu, Malaysia.

© 2011 Journal of Computing Press, NY, USA, ISSN 2151-9617



comp plete system. Without thorou analysis, it is impossibl to W ugh le have a good de esign or corr rect impleme entation [11]. In mode eling, analysis of the seman quality can help student in s ntic n ts desig gning the UML class diagram By doing semantic analy L m. s ysis, we de elve even deep to check whether they fo per w orm a sensible set e of in nstruction in the programm t ming language This will help e. h stude ents to overco ome the variation of nam ming the notation eleme because it is semanticall valid. ent ly

Wor rdNet ver rsion
2.1 2 3.0 3

Noun #N Syn nsets
81, ,426 82,115

erb #Ve Synsets
13,6 650 13,7 767

117,0 097 117,7 798

#Verb b
11,488 8 11,529 9

2.1 R RiTa.WordNet Background
RiTa covers a rang of computa ge ational tasks related to liter r rary pract tice, including text analy ysis, generatio on, display and anim mation, text-to-speech, text-m mining, and access to external a resou urces (e.g., Wo ordNet) [12]. RiTa.WordNe (RiWordNet) is et one o the core ob of bjects in RiTa toolkit that su upport structu ures for a specific task. Briefly, RiWo ordNet is an intuitive interface i to the WordNet on e ntology provid ding definition glosses, an a ns, nd range of -onym e ms (hyperny yms, hypony yms, synony yms, anton nyms, merony yms, etc.). It can be transp parently bund dled into a web-based, browser-exe , ecutable prog gram. RiWord dNet used for acces ssing WordNe via the RiTa et aServer. For most m also u cases it is simpler just to use the RiWordNet. s, j e

age of Word dNet such a Chinese W as WordNet, Th hai langua WordN Net, EuroWo ordNet and many more To improve e. WordN and its wi Net idespread app plications, ther is still a lot of re works to do. Each v versions of W WordNet noted the increased er ynsets, VerbSy ynset, Noun and Verb. Th his numbe of NounSy proves that WordNe continues to grow which is an advantage s et o online databas In Fig. 1, th escalation o the number of se. he of of an o

2.2 W WordNet Conc cept
Word dNet is a large lexical datab e base of (any) language. Nou l uns, verbs adjectives an adverbs are grouped into sets of cogni s, nd e o itive synon nyms (synsets each expres s), ssing a distinct concept. Syn t nsets are in nterlinked by means of co y onceptual-sem mantic and lex xical relati ions. The Wo ordNet is a sy ystem to integrate lexical and sema antic relations between the words. It org ganizes the lex xical inform mation in term of word me ms eanings and can be termed as a lexico based on ps on sycholinguisti principles. ic Inh herent to the technology im mprovement and realization of a n the im mportance of semantic, Wor rdNet has bee developed as a en great tool which gi t ives a great im mpact to the educational wo e orld espec cially in the semantic fi ield [13]. Few studies were w publi ished on using WordNe for education [14] or in et r specialized domai ins [15]. The are a lot of studies for ey t ucing variety languages of WordNet such as Ma y alay produ Word dNet [16], Th WordNet [17], EuroW hai WordNet [18] and many more. This shows that WordNet has been accepted all y W b d over the world. WordNet has been released in some versio W b i ons. y umber of wo ords. Every new version shows the increasing nu Table 1 and Table 2 show the st e tatistical overv view of Word dNet from version 1.6 to 3.0. o

synsets and words c be seen cl s can learly from W WordNet versio on 1.6 to v version 3.0.

Fig. 1. Es scalation of WordN synsets and words Net


WordNet Se earch Results

Wo ordNet ve ersion
1.6 1.7 2.0

#N Noun Synsets
66 6,025 75 5,084 79 9,689

#Ve erb Syns sets
12,1 127 13,2 214 13,5 508

#Nou un
94,47 74 109,195 114,64 48

#Verb b
10,319 9 11,088 8 11,306 6

The res sults of a sear of the Wor rch rdNet databas are displayed se in the R Results Wind ow. Horizontal and vertica scroll bars are al a t present for scrolling through the search resul All search g e lts. hes other t than overview list all senses matching th search results w he in the f following gene format. eral 1. Items enclos in italicize square brackets ([…]) may sed ed not be prese ent erformed on some senses of 2. If a search cannot be pe he ults searchstr, th search resu are headed by a string of the form: X of Y senses of searchstr. f 3. One line lis sting the num mber of sense matching the es search selected shows clearly the basic synt appear in W tax WordNet search Fig. 2 s results . Se ense n [sy ynset_offset}] [< <lex_filename> word1[#sen >] nse_number][, word2….] ,
Fig. 2. Ba syntax in Wor asic rdNet search resul lts

atistical overv view of Wor rdNet version are important ns Sta espec cially for those who are int terested in de eveloping another


Ba ased on Fig. 2, WordNet lexicographer files will app 2 pear when some optio n ons are select ted for extend ded view. Th hese lexico ographer files will be proce essed by grind which then will d produ uces a databas suitable for use with the WordNet libr se r rary, interf face code, and other applic d cations [20]. The format of the T lexico ographer files is described in wninput [2 A file num i 21]. mber corre esponds to ea ach lexicographer file. File numbers are F ded l W em cient encod in several parts of the WordNet syste as an effic way to indicate a lexicographe file name. By using the file er B ames, it can be used by prog e grams or end users to corre elate lexna file n names and num mbers [20]. As a case study of this rese s y earch, the rev view of Word dNet itself is very impor rtant. In this research, Word dNet will be used u to ex xtract its syns sets for object t-oriented mo odel. WordNe is et well-known softw ware used for many researches. The bination of the esaurus and dictionary mak it very us d kes seful comb espec cially in the semantic domain. One of the promin o nent exam mples of the us of WordNe is to determ se et mine the simila arity betwe words. Various algorit een V thms have bee proposed, and en these include considering the di e istance betwee the concep en ptual categ gories of word as well as considering the hierarch ds, a g hical struct ture of the WordNet on ntology. A number of th n hese Word dNet-based word similarity algorithms are implemen w y nted in a Perl package called Word dNet::Similarity and in Phy yton age TK packa called NLT [22]. WordNet Applic cations 2.4 W There are variou application of WordN e us ns Net especially in y comp puting domain Search has been run in a database that n. s n acces from Perpu ss ustakaan Nur Zahirah, Un r niversiti Malay ysia Teren ngganu databa ase, ScienceDi irect since 200 07-2011 as sho own in Fig 2. The keyw g. word used in th search is “w his wordnet” and the d date a access to this database is on 16th March 2011. d n 2

In nformation re etrieval and extraction e 1. .


Wo ordNet Applications
Describe a m method for the acquisition of th he hyponymy lex xical relation from unrestricted te m ext [23]. 2. As semantic lexicon in a m . module for full te ext message retr rieval in a commu unication aid [24]. 3. As a linguistic knowledge tool [25]. . c 4. As a tool f . for the automatic construction of thesauri [26], [27]. 5. Pearce, D. d . describes the use of WordNet in a e new techniqu for collocatio extraction [28 ue on 8]. They are: 6. Choueka [29 N-grams from 2 to 6 words in . 9]: m length. 7. Church and Hanks [30]: des . scribes techniques that used m mutual information to measure th n he strength of as ssociation betwee words. en 8. Smadja [31]: infer syntax b measuring th . by he spread of the distribution of c e counts between th he two collocates s. 9. Lin [32]: ba . ases his extraction method o on dependency triples obtained from a shallow d wparsed text co orpus. 10 Katz [33]: us patterns of parts-of-speech to 0. se extract technical terms (closely related to collocations). 1. Pedersen, T. et al. Compar 3 unsupervise . . re ed learning tha distinguish th sense of a at he an ambiguous w word in untagged t text [34].  McQuitty similarity analysis [35]. y’s  Ward’s m minimum-variance method [36].  EM algo orithm [37]. 2. The use of WordNet for improving search . f engines base on MORE techniques develope ed ed by RBSE pro oject for more e efficient retrieval of Internet resou urces [38]. 3. As a method . dological elemen to disambigua nt ate the meaning of words in info ormation extractio on systems [39]. 1. Resnik [40] computes class similarity, definin . ng class to be the nouns of a synsets plus th he nouns in all subordinate synse ets. 2. FEDDICT, a prototype onl . line dictionary to develop on IR technique base on measureme R ed ent of the concep ptual distance bet tween words [41]. 3. McTavish [42 used WordNet semantic domains . 2] to establish categories that could be used to analyze conc ceptual semantic distances in term ms of social env vironment to bett organize term ter ms for retrieval. 4. A. Budanitsk [43] classify t . ky that computation nal linguistics is the area that has placed th t he greatest emp phasis on relatio ons and semantic distances bet tween lexemes. 1. REBUILDER system is a first step to th . he development of a commercia CASE tool th al hat addresses th support of so he oftware design an nd design knowle edge manageme [44]. ent

Disa ambiguation

Semantic S distance

Object-oriented model

Fig. 3. Number of docu . uments published about WordNet in ScienceDirect (2 n 20072011)

Ba ased on Figure 3, WordN has been used widely in Net y Sema antic Web topi in 2007-2011 This is beca ic 1. ause the natur of re Word dNet consists of several semantic re s elations such as h synon nym, antonym hyponym an many more m, nd e. Ta able 3 discusse several rese es earches that have been don in h ne vario ous domains. Some popular domains are information retrie eval and extrac ction, disambi iguation and semantic distance. s We also show resea arch related to object-oriented model. o

This fr ramework con nsists of five phases. The d details of eve ery phase a discussed as follow. are Phase 1 hase require i input as stud dents’ answers which used to s This ph match with answers scheme prepa ared by lectur The answe rer. ers means UML cl lass name. If the class nam answered by me here m studen nts match th answers scheme, the answers are he e a conside ered to be true If not, it will continue to t 2nd phase. e. the .



Phase 2 Phase 2 is synsets extraction process in which synsets will be extracted from WordNet to match both answers. In this phase, we used RiTa.WordNet to get all the synsets for object-oriented model. After the extracting process, then the stemming process begins before produces the last synsets in the root form. The synsets extracted from this process will then be used in the next phase for the matching process. Phase 3 Phase 3 is the matching process. This process match the synsets extracted from the previous phase with the answers scheme used in the first phase. Phase 4 If one of the synsets extracted match the answers scheme, synsets then be stored as match word (class name) and updated in the knowledge-based. Phase 5 Finally, the output generated by the matching process. Phase by phase of the framework can be seen clearly in Fig. 4.
Phase 1: Input Students’ answers Match? Answers scheme

provide support for automated software testing [49], [50], [51]. The Testing framework is responsible for [52]: 1. Defining the format in which to express expectations 2. Creating a mechanism to hook into or drive the application under test 3. Executing the tests 4. Reporting results In this research, a tool is needed to prove the FASOOM framework. RiTa.WordNet as one of the core objects in RiTa tools has been chosen for this purpose.

In this research, synonyms are important to make sure that answers given by students are true. For the early stage, we design a simple program to extract synonym sets (synsets) using Rita.WordNet. Several words that usually used as class name to UML have been tested and the outputs are shown in Table 4.

Search words

Synsets 0: history Synsets 1: chronicle Synsets 2: story Synsets 3: report Synsets 4: explanation Synsets 5: score Synsets 6: accounting Synsets 7: bill Synsets 8: invoice Synsets 0: client Synsets 0: customer Synsets 1: node Synsets 2: guest Synsets 0: dealing Synsets 1: dealings No Synsets!



Phase 2: Synsets Extraction Process Phase 5: Output WordNet
customer client

Phase 4: Storing the matched synsets


Extraction using Rita.WordNet


Stemming process

Phase 3: Matching Process Answers scheme Matching method Synsets extracted

Initially, we test only five words that are commonly used for UML class name. They are account, customer, client, transaction and money. Customer and client are two words those synonyms to each other. We test these two words to see that if they can give the same output.

As shown in Table 4, a search word gives several synsets or no synsets at all. This provides several choices for students in naming the UML class diagram correctly even though they are not provided in answers scheme. RiTa.WordNet has been used to extract synsets from WordNet. Other extractor tools such as TextCatch [53], TextToOnto [54] and Email Extractor [55] have been reviewed and obviously RiTa.WordNet is the best choice. The used of RiWordNet are because of its significant in this research. We can conclude that RiWordNet are 1. Easy to use and understand 2. Access directly to WordNet database through RiTaServer. 3. Give the simple output as requested

Fig. 4. FASOOM Framework

A lot of research has been done on object-oriented programs testing. However, few of them address the problems related with integration testing [45], [46], [47]. In other research, tool has been developed for testing object-oriented programs at the integration level, accomplishing the automation of both test case generation and test execution [48]. A test automation framework is a set of assumptions, concepts or tools that




Easy to customize programming)





[6] [7]

Even though RiTa.WordNet gives a lot of benefits compared to other tools, some problems appear when we search words in the same synsets. As an example in Table 4, when we search for the word “customer”, only client has appeared to be the synsets for customer. Otherwise, when the search for the word “client” is made, there are two other synsets found such as “guest” and “node”. If client is a synonym for customer and guest is a synonym for the client, we can say that the word guest is also synonym to customer. However, the results that we have gained show that RiTa.WordNet cannot give the overall synsets for the search words. Therefore, some modification should be made to retrieve all the synsets of the search word.






The outcome from this study is a new framework introduced to semantically analyze the object-oriented model. To carry out this framework, a tool called RiTa.WordNet is adapted to get synsets from WordNet and the synsets then matched with the input given by the user. The extract and match step are tremendously pertinent to analyze the semantic in the proposed framework. By this development, users do not need to search their synonym information directly from WordNet. This framework development helps students especially in their searching of related meaning, in the specific domain (especially in the modeling domain). In addition, students will save time on completing their tasks with accurate outcome. We believe that this study will give a significant impact in adaptation of WordNet in the other way. Future work will focus on gaining more synsets that commonly used for UML class name using Rita.WordNet. For this stage, we will find the solution to get more synsets for the search words.

[11] [12] [13] [14]




This research was supported by a grant Tabung Bantuan Pendidikan Khas (TBPK), Universiti Malaysia Terengganu (Vot: 53057) and National Science Fellowship (NSF) under Ministry of Science, Technology and Innovation (MOSTI) Malaysia. We would like to express our thanks to all people which helped in completing this paper.
[18] [19] [20] [21] [22]

[1] [2] Johnson, R.E. and B. Foote, "Designing reusable classes," J. ObjectOriented Programming 1, 5, vol. no. pp. 22-35, 1988 Fayad, M.E., D.C. Schmidt, and R.E. Johnson, "Object-Oriented Application Frameworks: Problems and Perspectives," vol. no. pp. 1997 Riehle, D., Framework Design: A Role Modeling Approach. 2000: ETH Zürich. Teeuw, W.B. and H.v.d. Berg, On the Quality of Conceptual Models. 1997. Ali, N.H., Z. Shukur, and S. Idris, "Pendekatan Formalisasi Model Berorientasi Objek dengan Model Formal: Satu Tinjauan," Jurnal


[3] [4] [5]



Teknologi, vol. 49, no. D, pp. 1–12, 2008 Booch, G., "UML in action," Commun. ACM, vol. 42, no. 10, pp. 26-28, 1999 Christian, F.J.L., Improving the quality of UML models in practice, in Proceedings of the 28th international conference on Software engineering. 2006, ACM: Shanghai, China. Thomasson, B., M. Ratcliffe, and L. Thomas, Identifying novice difficulties in object oriented design, in Proceedings of the 11th annual SIGCSE conference on Innovation and technology in computer science education. 2006, ACM: Bologna, Italy. Noor Maizura Mohamad Noor, Noraida Haji Ali, and Noor Syakirah Ibrahim, "A New Framework to Extract WordNet Lexicographer Files for Semi-Formal Notation: A Preliminary Study," Proc. 4th Internationatiol Symposium on Information Technology 2010 (ITSim'10), pp. 1027-1031, 2010. Nor Fazlida Mohd, S., Z. Abdullah Mohd, I. Sufian, and S. Zarina, Designing an understanding and debugging tool (UDT) for objectoriented programming language, in Proceedings of the 4th WSEAS International Conference on Artificial Intelligence, Knowledge Engineering Data Bases. 2005, World Scientific and Engineering Academy and Society (WSEAS): Salzburg, Austria. Eduardo, F. and Y. Xiaohong, Semantic Analysis Patterns, in Conceptual Modeling — ER 2000. 2000. pp. 918-924. Howe, D.C., RiTa: Creativity Support for Computational Literature, in SIGCHI 2008. 2008, ACM: Florence, Italy. pp. 4. Miller, G.A., "WordNet: A Lexical Database for English," Communication of the ACM, vol. 38, no. 11, pp. 39-41, 1995 Kegl, J., Machine-readable dictionaries and education, in automating the lexicon: research and practice in a multilingual environment, D. Walker, et al., Editors. 1995, Oxford University Press: Oxford, New York. pp. 249–284. Strzalkowski, T. and R. Brandow, "Spotting technical concepts in natural language text," Proc. Proceedings of the 9th Florida Artificial Intelligence Research Symposium, pp. 66-70, 1996. Tze, L.L. and N. Hussein, Fast Prototyping of a Malay WordNet System, in Proceedings of the Language, Artificial Intelligence and Computer Science for Natural Language Processing Applications (LAICS-NLP) Summer School Workshop. 2006: Bangkok, Thailand. pp. 13–16. Sareewan, T., R. Kergrit, M. Chumpol, S. Tan, C. Thatsanee, S. Virach, and I. Hitoshi, Thai WordNet construction, in Proceedings of the 7th Workshop on Asian Language Resources. 2009, Association for Computational Linguistics: Suntec, Singapore. Piek, V., "Introduction to EuroWordNet," Computers and the Humanities, vol. 32, no. 2, pp. 73-89, 1998 Yu, J., Z. Wen, Y. Liu, and Z. Jin, "Statistical Overview of WordNet from 1.6 to 2.0," Proc. Global WordNet Conference, pp. 352-357, 2004. Princeton, "lexnames (5WN) manual page," vol. no. pp. 1-3, 2005 Princeton, "wninput (5WN) manual page," 2005 Ted, P., P. Siddharth, and M. Jason, WordNet::Similarity: measuring the relatedness of concepts, in Demonstration Papers at HLT-NAACL 2004 on XX. 2004, Association for Computational Linguistics: Boston, Massachusetts. Hearst, M.A., Automatic acquisition of hyponyms from large text corpora, in Proceedings of the 14th conference on Computational linguistics - Volume 2. 1992, Association for Computational Linguistics: Nantes, France. Van de Riet, R., H. Burg, and F. Dehne, "Lingusitic instruments in information system design. FOIS," Proc. Proceedings of 1st International Conference, 1998. Jeong-Oog, L. and B. Doo-Kwon, Semantic Integration of Databases








[31] [32] [33]







[40] [41]




Using Linguistic Knowledge, in Proceedings of the 12th Australian Joint Conference on Artificial Intelligence: Advanced Topics in Artificial Intelligence. 1999, Springer-Verlag. Dongqiang, Y. and M.P. David, Automatic thesaurus construction, in Proceedings of the thirty-first Australasian conference on Computer science - Volume 74. 2008, Australian Computer Society, Inc.: Wollongong, Australia. Mandala, R., T. Tokunaga, H. Tanaka, A. Okumura, and K. Satoh, Ad Hoc Retrieval Experiments Using WordNet and Automatically Constructed Thesauri., in TREC-7. 1998. pp. 414–419. Pearce, D., "Synonymy in collocation extraction," Proceedings of the NAACL 2001 Workshop on WordNet and Other Lexical Resources: Applications, Extensions and Customizations, vol. no. pp. 41-46, 2001 Choueka, Y., "Looking for needles in a haystack or locating interesting collocational expressions in large textual databases.," Proc. Proceedings of the RIAO '88 Conference on User-Oriented ContentBased Text and Image Handling, pp. 1-15, 1988. Church, K.W. and P. Hanks, "Word association norms, mutual information and lexicography," Computational Linguistics, vol. 16, no. 1, pp. 22-29, 1990 Smadja, F., "Retrieving collocations from text: Xtract," Computational Linguistics, vol. 19, no. 1, pp. 143-177, 1993 Lin, D., "Extracting collocations from text corpora," Proc. First Workshop on Computational Terminology, pp. 1998. Justeson, J.S. and S.M. Katz, "Technical terminology: some linguistic properties and an algorithm for identification in text," Natural Language Engineering, vol. 1, no. 1, pp. 9-27, 1995 Pedersen, T. and R. Bruce, "Distinguishing Word Senses in Untagged Text," Proc. Second Conference on Empirical Methods in Natural Language Processing, 1997. McQuitty, L., "Similarity analysis by reciprocal pairs for discrete and continuous data," Educational and Psychological Measurement,, vol. 26, no. pp. 825-831, 1966 Ward, J., "Hierarchical grouping to optimize an objective function," Journal of the American Statistical Association, vol. 58, no. pp. 236– 244, 1963 Dempster, A.P., N.M. Laird, and D.B. Rubin., "Maximum likelihood from incomplete data via the EM algorithm," Journal of the Royal Statistical Society B, vol. 39, no. pp. 1–38, 1977 Eichmann, D., Balancing the Need for Knowledge and Nimbleness in Transportable Agents. Position paper for the Dartmouth Workshop on Transportable Agents, 1996. Chai, J.Y. and A.W. Biermann, A WordNet Based Rule Generalization Engine for Meaning Extraction System, in Proceedings of the 10th International Symposium on Foundations of Intelligent Systems. 1997, Springer-Verlag. Resnik, P.S., Selection and information: a class-based approach to lexical relationships. 1993, University of Pennsylvania. Richardson, R., A.F. Smeaton, and J. Murphy, Using WordNet for conceptual distance measurement., in Proceedings of the 16th Research Colloquium of the British Computer Society Information Retrieval Specialist Group. 1996, Taylor Graham: London. pp. 100123. McTavish, D.G., K.C. Litkowski, and S. Schrader, "A computer content analysis approach to measuring social distance in residential organizations for older people," Social Science Computer Review, vol. 15, no. 2, pp. 170-180, 1997 Budanitsky, A. and G. Hirst, "Semantic distance in WordNet; An experimental, application-oriented evaluation of five measures," vol. no. pp. 2001 Gomes, P.J.d.S., A Case-Based Approach to Software Design, in Departamento de Engenharia Inform´atica. 2003, Universidade de

Coimbra: Coimbra. pp. 284. [45] Chen, H.Y., "An approach for object-oriented cluster-level tests based on UML," Proc. Proceedings of the IEEE International Conference on Systems, Man and Cybernetics, pp. 1064-1068, 2003. [46] Basanieri, F. and A. Bertolino, "A Practical Approach to UML-based Derivation of Integration Tests," Proc. 4th International Software Quality Week Europe, pp. 2000. [47] Hartmann, J., C. Imoberdorf, and M. Meisinger, "UML-based integration testing," Proc. Proceedings of the ACM SIGSOFT 2000 International Symposium on Software Testing and Analysis, pp. 6070, 2000. [48] Tom, M. and L. Zhe, A test framework for integration testing of object-oriented programs, in Proceedings of the 2007 conference of the center for advanced studies on Collaborative research. 2007, ACM: Richmond Hill, Ontario, Canada. [49] Eun Ha, K., N. Jong Chae, and R. Seok Moon, Test Automation Framework for Implementing Continuous Integration, in Proceedings of the 2009 Sixth International Conference on Information Technology: New Generations. 2009, IEEE Computer Society. [50] Rankin, C., "The software testing automation framework," IBM Syst. J., vol. 41, no. 1, pp. 126-139, 2002 [51] "Test Automation Framework," [52] "Automation Framework," [53] Skesoft, "Text Catch 2.7.12," 2008 [54] Geeknet, "TextToOnto," [55] WebPro, "Email Extractor 1.4," 2009 Noraida Haji Ali received her Bachelor and Master degree in Computer Science from Universiti Kebangsaan Malaysia in 1995 and 1999 respectively. She is currently a lecturer in Universiti Malaysia Terengganu and is doing her PhD study at Universiti Kebangsaan Malaysia. Her research interests include Obejct-Oriented Modeling and Formal Method. Noor Syakirah Ibrahim obtained her Bachelor degree in Information Science from Universiti Kebangsaan Malaysia in 2009. She is currently doing a Master study in Universiti Malaysia Terengganu. Nor Fazlida Mohd Sani is a senior lecturer in the Department of Computer Science, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia. She obtained her Ph.D. focus on Program Analysis and Program Understanding System from Universiti Kebangsaan Malaysia in 2007. Her research interests include reverse engineering, software maintenance, program debugging, software security and database. Noor Maizura Mohamad Noor received her Diploma and Bachelor Degree in Computer Science Universiti Pertanian Malaysia in 1990 and 1994 respectively. In 1997, she obtained Master Degree in Computer Science from Universiti Putra Malaysia. She also received her phD Degree from The Manchester University in 2005. She is currently a senior lecturer at Universiti Malaysia Terengganu and interest in research including Information Systems, Decision Support System and Clinical Decision Support System.