component) based on methodologies for Ontology Learning (OL). In a. University case, we had applied a Systemic Methodology for OL (SMOL) from a RDB to ...
An Ontology-Learning Knowledge Support System to Keep e-Organization's Knowledge Up-To-Date: A University Case Study
R.J. Gil and M.J. Martín-Bautista
Dept. of Computer Science and Arti cial Intelligence, University of Granada, Granada, Spain
e-Organizational users can apply semantic engineering solutions to deal with decision-making and task-intensive knowledge requirements supported by Knowledge Management Systems (KMSs). Such optional engineering strategies consider some system types to meet knowledge users' need, aligned with the e-services and e-management qualities required for them. Particularly, in the Knowledge Support System (KSS) eld, developers have adopted some Ontology-based technologies to support user's task-knowledge system functionalities. In this paper, an Ontology-Learning Knowledge Support System (OLeKSS ) model is proposed as a general component of e-organizations, to keep the ontologies associated with this kind of KMS updated and enriched. Relational Databases (RDBs) are considered complementary knowledge source for Knowledge Acquisition (KA) through a OLeKSS Process (as a subsystem component) based on methodologies for Ontology Learning (OL). In a University case, we had applied a Systemic Methodology for OL (SMOL) from a RDB to update the correspondent host-ontology associated to the University's KSS during this OLeKSS process. Abstract.
1
Introduction
According to Finquelievich and Prince [18] the e-University is the intensive, extensive and strategic application of the novel Information, Communication and Internet Technologies (ICTs) to every university's activities; particularly, when these e-Universities are trying to be competitive and more e cient trough their innovative information systems implementations [37][45]. Indeed, under the Knowledge Management (KM) umbrella, modern organizations have applied several alternative technologies for Knowledge Acquisition (KA) and processing from speci c domains associated with the information systems of e-organizations such as e-Universities [7][52]. In accordance with the de nition given in [48]; KM is a systematic method for managing individual, group and organizational knowledge using the appropriate means and technology. At its root it is to do with managing people, what they know, their social interactions in performing task, their decision making, the way information ows and the enterprise work culture.
2 For our proposal, we must point out three relevant aspects of KM: a) Collectivist knowledge perspective: The social aspect of the knowledge (groups, communities, networks, and organization units) are considered rather than the individualist one [51]. b) Knowledge propertie s for re-usability : on one hand, the capability of knowledge to generate new information as a product of the intervention and processing carried out by users, and on the other hand, the implicit and explicit quality features and the possibility of converting from the former to the later [39], and c) Knowledge as a competitive resource from heterogeneous electronic knowledge sources (e-KSOs): First, knowledge is an assessment for decision management support. Likewise, knowledge can be used as technical representation methods such as ontologies, knowledge artefacts, etc, and, nally, dierent e-KSOs can be learned through diverse Methodological Resources (MRs) [47]. Knowledge Management System (KMS) includes Knowledge Based Systems (KBS) and Knowledge Support System (KSS) [43]. The former ones, can be seen as a type of KBS following the collectivist perspective of KM, the later ones are oriented to assist knowledge activities such as organizational practices and routines (e.g. Document management), knowledge distribution (e.g. Groupware), and knowledge adoption purpose (New products and markets) [6]. However, KBS may qualify better as software developed to satisfy speci c user's needs usually as an application for expert decision-making [4]. Gaines [20] establishes that KSSs must be speci ed in terms of their roles in the social knowledge process. This is in contrast to specify them in technical or individual cognitive terms, as it is common currently. This author suggests that there are some important requirements that a KSS must satisfy; starting by a overall requirement: A KBS is a component in a structure of social knowledge process. Specify the top-level requirements in terms of the relevant structure, existing knowledge processes, intended changes and unexpected side-eects. And three meta-requirements as well: 1) The social structure usually involves a professional community responsible for managing the processes of knowledge acquisition and dissemination and a client -users- community dependent on the knowledge for its activities. 2) A KSS will not contain all the knowledge relating to the processes in which it is involved, or provide all the facilities required. And, 3) Knowledge processes are intrinsically re exive, applying to themselves. On the other hand, regarding the dierent e-KSOs used by e-Organizations for KA such as texts, databases and even other ontologies previously developed for this domain, we have centered the interest in KSS's types which are based their Ontology Learning (OL) processes from Relational Database (RDB). In this sense, for the study case the recent technical literature about the diverse options available for retrieving knowledge from RDB sources has been reviewed for the study case. This includes the RDB models and the schemes, as well as the knowledge that underlies data which can be retrieved and processed using data mining techniques. Some RDB conversion-to-ontology tools have been analyzed, trying to design a strategy compatible with the methodology applied to the case.
3 The principal concern for this work is: 1) Reviewing the KSSs under a social perspective trying to increase the eKSO available to keep them as functional and updated as users require. 2) Exploring the Methodological Resources (MRs) such as methods, techniques and tools available for knowledge updating and enrichment processes of the KSS's semantic structures (e.g. ontologies) through OL from diverse e-KSOs. Particularly, the ones related with OL from Relational Databases (RDB). 3) Modelling under the systemic approach, some KSS's pro les which are developed using ontological mechanism as part of their system architecture and implementations. The main contributions in this work are: 1) A new perspective of KSS based in OL from diverse e-KSOs is suggested considering the previously exchange social-relationship de ned between knowledge generation communities (experts) and the communities of users [20] trying to take advantage of the e-society interactions. 2) Identifying the common KSSs characteristics for functional systems help to conceive a general system architecture (main components) that could be drafting a model which should be improved through ontological engineering. 3) A Systemic proposal for Ontology Leaning KSS (OLeKSS) model has been conceived and speci ed showing the relationships among components. 4) The Systemic Methodology for OL (SMOL) is applied for an academic case study to enhance a speci c ontology-based KSS through OL from RDB. Although these contributions are reached in this work, some of them should be summarized in the following sections to meet the pages limit established. This article is structured as follows: Ontology Learning as a methodological option is reviewing in Section 2. A new approach of KSS based in ontologies is shown in Section 3. The model speci cation of KSS based in OL named OLeKSS is suggested in Section 4. A short description of SMOL is explained in section 5. The application of SMOL to a real case study is shown in Section 6. Finally, the conclusions are presented in Section 7.
2
Ontology Learning as a Methodological Option
Several OL de nitions related to methodologies approach are given in the technical literature [26][27]. They are usually relative to some methods and techniques used to improve ontologies previously developed from a speci c e-KSO. Some MRs for OL involved with texts or documents collected in a corpus is the most common e-KSO referred in the literature [19]. We have been extended this partial OL perspective under a novel model of ontology-based KSS which may include dynamic and continuous learning processes as an essential system component, considering diverse MRs from heterogeneous and complementary e-KSOs: texts [8][12][40], ontologies [15][16][42], and databases [5][11] allowing to use of more than one e-KSO in the same model. Thus, the role of the users in a OL methodology for KSS improvement, it is to combine dierent MRs for diverse e-KSOs in an appropriated way, consid-
4 ering the existence of one or more domain ontologies already elaborated for a KSS which could be enhanced through updating and enrichment OL processes [28][38]. These e-KSOs are grouped according to our proposal as follow: a) OL from other ontologies developed previously, b) OL from content of texts or compendiums of them, and c) OL from database schemes and their data-values. A methodological proposal for us named SMOL [21] is suggested by us as an option for KA processes into the KSS, because it considers the three cited e-KSOs. This methodology is described shortly in subsection 5.
2.1
Methodological Resources Useful for Ontology Learning
There are some de nitions regarding to MRs that allows us to understand and clarify the concepts associated with them, usually confused in technical literature. The following de nitions and the performance measures associated to them have been considered [9]:
Techniques: subjective capabilities (abilities or skills) to handle a tool properly. Performance measure: e ciency. Methods: a way of thinking or doing using a tool to achieve an objective. Performance measure: eectiveness. Tools: objective capabilities to use the resources properly to apply techniques. Performance measure: e ciency. Methodologies: a related set of methods, techniques, and tools which could be used to reach objectives. Performance measure: e cacious (eectiveness/e ciency).
Some representatives MRs for OL are summarized according to each e-KSO in [24]. We have use some of them in SMOL applied to the case study in Section 6.
3
Knowledge Support Systems based on Ontologies
In the literature, we can nd some approaches of ontology-based KSSs for dierent purposes. These types of KSSs are developed to satisfy user's requirements taking advantage of ontology framework to represent knowledge structures and rules associated. Some common representations are related to: documents, user's pro les, task -pro les and -work ow, agent's coordination, and so on. Recently, some ontology-based KSSs have been developed for knowledge sharing, -task-based, -collaborative and -recommender purposes [1]. They are reviewed and grouped below (Table 1) summarizing some recent works for each type of KSS, trying to nd common characteristics to suggest a general KSS framework as representative model [17].
3.1
Common Ontology-based KSS Characteristics
As we can see in Table 1 about the characteristics of the KSSs, there are relevant components such as: a) Linking users with their requirements, b) communications and sharing connections, c) processes for knowledge discovering and restructuring, and d) obtaining of knowledge products.
Table 1.
KSS's Pro le Knowledge Sharing (OntoShare System) Document Recovery (MILK System)
Task and Work ow (Liu & Wu) Context-aware and Processesaware (KnowMore & FRODOTaskMan systems). Problemsolving and Recommender Systems (RS)
Some Ontology-based KSS characteristics
Characteristics - Automatic Knowledge Sharing with aid of user's pro les (topicsconcepts). - Ontological concepts according user's interest. - Documents are represented as ontologies. - The explicit knowledge is recovered by e-mail, keywords, and documents. Implicit knowledge may be shared through user's pro les. - Distributed knowledge, located in dierent places should be integrated, - Cross-fertilization and communication should be supported among users, - Implicit organizational-members' knowledge should be accessible together with the explicit one. - Documents should be presented to users where & when they may need them. - Collective task-based workplace simplifying the knowledge retrieval and sharing among peer-groups - Task pro les to support knowledge workers - Information retrieval and ltering techniques for text-processing, indexing, querying and pro le tasks. - Heuristic ontology-based techniques to support task-work ow management. - KnowMore was developed to extend support to knowledgeintensive tasks, considering three key elements such as: 1) Information needs, 2) Context-aware, and 3) Ontologies (work ow- and domain-context - Information space (system component): use ontologies metamodels and document indexed under task pro les. - FRODO is conceived as an Agent Society based in ontologies. - User pro ling within RS is used to recommend on-line academic research papers. - RS that allows to customize content to be suggested based on the user's browsing pro le. - One of them, developed a novel task-based knowledge RS. - A work ow ontology-based according to the correlation among users, roles, and tasks.
5 Authors Davies et al [14] Lee et al [33]
Agostini et al [3] Jung et al [32]
Liu & Wu [34] Liu et al [35]
Abecker et al [2] Holz et al [29]
Middleton et al [36] Liu & Wu [34] Zhen et al [53]
In this sense, these KSSs are developed to oer rst, useful knowledge to Users according to their user's task and responsibilities, styles and preferences (contexts and pro les). Second, the KSS must warrant the dynamic Communication interchange concerned with user's task activities. Such task activities are increased among users as they require specialized knowledge for eective decision-making, creation, acquisition and identi cation. And third, these facts imply new knowledge searching, recovery and discovery Processes to achieve under an adequate performance, the structured knowledge required as semantic Products (e.g. ontology or context). These highlighted words are key elements considered by us as common present characteristics of this ontology-based KSS. Our proposal ful lls these requirements, allowing the ontology-based KSS to enable for knowledge -acquisition, -repositories, -discovery, and -distribution. We consider the improvement of the KM ability and their whole learning capability through OL. Moreover, we also consider dierent and complementary e-KSOs.
6 EXPERT USERS OL from Ontologies and Databases
Study Cases Examples/Cases Instruction Education Systems Theories
Apply Model
Form Model
Conferences Workshops Journals
Strategies
Collect Data
Data Elicit
Knowledge Engineering
Models & Ontologies
Texts OL from Texts
END USERS
Instancies & data
Discuss Evaluate Advice Advice
Experiences Problems Novelty
Needs & Requeriments Criticiam Resources Rewards
Advice
INTERNET
INTERNET
Figure 1: The social context for Ontology Learning knowledge support process Fig. 1.
3.2
The e-social context for Ontology Learning knowledge support process
KSS Improvement Through OL using Dierent e-KSOs
New knowledge is used and required for continuous KSS's updating process and this knowledge is represented as ontologies, despite this, e-KSO are omitted as main features or components of the KSSs. Indeed, e-KSOs such as texts, ontologies, and databases that could be used for semantic learning purposes, they are usually not explicitly cited by authors as key elements integrated in the KSS's architectures reviewed above. Moreover, we suggest OL process for Knowledge Acquisition (KA) [28] as a useful option to integrate these resources in the KSS architecture. The reviewed KSSs showed us how some ontology-based mechanisms have been applied to support user's knowledge task and work- ow requirements. Predominantly, ontologies developed for KSSs are used to represent and to describe e-organizational user's tasks and roles. Others represent task-work ow and content-structure (documents). Also, there are ontologies to support userpro les and context of use, including knowledge about collaborative relationships that may emerge among diverse organizational members.
3.3
Social Context of KSS based on Ontology Learning
In the Gaines' proposal, the KSS model works in a social context, representing, and describing the knowledge relationship exchange between professional communities (Expert role) and (the End-) user communities [20]. We propose a new e-social model (Figure 1) as an attempt to emphasize the OL potential for enhancing the Model and the Ontologies associated with the KSS using those possible e-KSOs (in dash-ovals). The communities' exchange mechanisms cited are related with the generation, transfer, assimilation, and re-conversion of knowledge in terms of the interchange relationship between the social communities that manage the knowledge (Experts) and those who use it (End-users). This knowledge exchange and interactive potential is reality possible based on the modern e-society. Through OL processes, the available knowledge gathered on previous case studies expressed as e-published ontologies (upper or domain), would be useful for enhancing the contextual-ontologies of KSS by comparing feed-forward.
7 OLeKSS USERS
Interface
Interface
OLeKSS PROCESSES
OLeKSS KNOWLEDGE SOURCES
Interface
OLeKSS PRODUCTS Structured knowledge
Methodological Methods Tools resource
(Specific ontologies) Contexts Agents & Knowledge Artifact
Techniques
OLeKSS COMMUNICATION
Figure 2: OLeKSS model component specification
Fig. 2.
OLeKSS model speci cation
Some knowledge needed for intensive tasks that users are developing in their study or system domain, require eective knowledge access and processing from content in texts. Through OL from texts, representative ontologies of this KSS would be updated whether the recoverable knowledge can be learned from scienti c texts (retrieved by the Internet) about conferences, workshops and journals. Similarly, available knowledge from equivalent or contrasting information systems databases (same e-organization or e-others) could be gathered by OL, to enhance the knowledge expressed in ontologies associated to the intended KSS.
4
An Ontology-Learning KSS (OLeKSS)
In this section, we present our proposal for a model of an Ontology-Learning KSS (OLeKSS). For a best understanding of the description of the model, we have used a process-based scheme coming from the eld of System Engineering [10][31]. Under this systemic approach of top-down abstraction-levels, based on the KSS meta-requirements suggested by Gaines [20], the components speci cation of the model are explained in the following subsection.
4.1
System Components of an Ontology-Learning KSS
The main OLeKSS model components are: Users, Processes, Products, Communications and Knowledge Sources. They are shown as ovals in Figure 2, represented in UML as Class diagram in Figure 4, and described as follows: OLeKSS Users obtain added value from OLeKSS Processes. They make decisions about the knowledge domain that they already have or that they are constructing from possible OLeKSS Knowledge Sources. Graphic User Interfaces may include necessary and ergonomic operational options that can simplify knowledge processing and visualization. It should include e cient options allowing to recover and to update of related OLeKSS Products. OLeKSS Processes are applications of a set of MRs such as methods, techniques, tools, and agents, with capability to construct or to update knowledge structures such as ontologies and other representation types. These processes
8 may enrich and adapt the existing knowledge in (semi) automatic way using and distributing information from heterogeneous OLeKSS Knowledge Sources. Thus, the resources needed during the developing period of time are reduced. OLeKSS Products are de ned based on partial results obtained during the OLeKSS Processes as well as on the structured or unstructured knowledge acquired previously (e.g. ontologies or pro les). Some particular results such as KSS subsystems (e.g. reusable Agents) are also considered as partial OLeKSS Products (e.g. Knowledge Artifact). Consequently, this fact will facilitate keeping those partial results accessible and updated as OLeKSS Knowledge Sources for any other OL reusing purpose. OLeKSS Communication supports internal and external communications among OLeKSS Users for sharing and transferring knowledge, to guarantee collaboration and coordination. Other connections needed to create and manage Expert-users knowledge networks about OLeKSS Knowledge Sources are considered. OLeKSS Knowledge Sources are dierent structured or unstructured sources that provide quali ed knowledge to sustain the sub-processes involved in the OLeKSS Processes. These sources may be useful for OLeKSS Users to gain easy knowledge access and processing mechanisms through storage catalogues or repositories. This mechanism may support e cient quality cycles about the users' versions- and updating- revision during OLeKSS Processes. A more detailed speci cation based on Classes under the UML approach can be found in Figure 4.
4.2
Relevant Aspects of OLeKSS
As we can see above, our proposal consider the OL from dierent e-KSOs would be improving the KSSs through knowledge growth, restructuring, and comparing processes related to: a) Knowledge-bases within those KSSs. b) Operational knowledge structures. And c) Structured ltering of resources. We could summarize some advantages related to this systemic approach as follow: a) It is showed explicitly how these e-KSOs (ontologies, texts and databases) could be associated with OLeKSS through OL mechanisms based on the e-society possibilities (Figure 1). b) The OLeKSS architecture (Figure 2) represents pro les of KSS developed previously but consider an important systemic component also: the knowledge sources and the Methodological Resources (MRs) involved and expressed in OL methodologies. And, c) the exibility of our design shows how the knowledge associated to the OLeKSS could be recovered and updated from e-KSOs applying any accredited OL methodology. On the other hand, there are not previous references about KSS's architectures considering the methodologies and e-KSOs as systemic components as well. Regardless of the characteristics of this type of KSS requiring continuous KA processes, it is more common to nd references dealing with the OL process considering each e-KSO for separated. We have found only two recent references that consider not more than two KSOs cited for speci c KA processes, independently of the supported or involved system [41][49].
9 c
III- Query requirements
User’s Profiles
(Protégé)
Users
a
(Protégé)
IV- Knowledge selection
V- Knowledge structure construction
Access & MySQL
RDB-IUTEPAS
RDBToOnto & ODEMapster
a
Fig. 3.
5
VII- Knowledge structure reorganization
Versioning
II- Knowledge Discovery
I- Methodology strategy selection
University Database (RDB-IUTEPAS) Selected strategy
b
Satisfied requirement
b
VI- Knowledge exploring and searching
Knowledge Sources c
Knowledge structure updating
Semantic Products
DEAOntology updated Phase Decision point Phase-Flow Data flow Database
1 Application of SMOL methodology to the case study
The OLeKSS Processes Supported by SMOL
Although any accredited methodologies such as DynamOnt, KACTUS, ONIONS, On-To-Knowledge, DINO, SENSUS, and Simperl's et al. proposal [13][49] could be used to support of OLeKSS Processes, in this work we applied a Systemic Methodology for Ontology Learning (SMOL) developed by us [24][25]. Particularly, SMOL tries to conciliate the system total quality paradigms with usercentered services to meet user's requirements [10]. Likewise, it combines some MRs previously developed according to the available e-KSO. SMOL was selected for the following reasons: a) It was developed under the systemic approach also; b) it considers any e-KSOs as expected sources; and c) it was tested for other cases (ontologies/texts) in previous works [22][24]. Some other SMOL's advantages are: user-oriented, integrated, exible, open, interactive, and iterative. SMOL's ow is showed in Figure 3 and brie y described (due to space limitation) as follow: I. Methodology strategy selection. The complexity of the domain is evaluated. II. Knowledge discovery. The MR from dierent e-KSOs and repositories are combined. III. Query requirements. Dierent queries are formulated to the knowledge sources available by browsers or another kind of applications. IV. Knowledge selection. A selection of the retrieved information from the formulated queries to the e-KSOs and repositories is performed. V. Knowledge structure construction. Dierent structures such as ontologies and contexts can be built interactively with users' advisory. VI. Knowledge exploring and searching. The knowledge structures are explored, veri ed, and validated. VII. Knowledge structures reorganization. Processes such as grouping of instances, ontology population and other activities are similarly performed. And VIII. OLeKSS Con guration. Users set up modules of the OLeKSS associated with ontologies. Decision points have been included for the participation of the user in the checking of the quality. Some of them are shown in Figure 3 in rhombus shapes.
6
Academic Case Study
A speci c University case study has been selected as an experimental academic domain to test the OLeKSS model proposal under a methodological focus [44].
10
Table 2.
Conversion tools according to capability parameters [46]
Conversion Tool
Mapping Mapping Mapping im- Query imple- Application Data creation representation plementation mentation domain integration DataMaster Automatic Logic rules Static Generic/ Possible Speci c (Protégé) RDBToOnto Automatic Constrain Without Potential Generic None rules (user explicit SPARQL interaction) (data-mining) registration ontology population ODEMapster Both R2O both SPARQL -> Generic Possible multiple (auto- & by Language (static & RDF/SQL -> sources user) dynamic) RDB. The host-ontology (called DEA-Ontology) selected to be updated and enriched was obtained from an Ontology Development process from an expert support system. It was designed to provide study option recommendations for students matriculated in a University of Venezuela that operates under a Distance Education Administration (DEA) [23][50]. The main University processes supported with this OLeKSS are admission and distance-courses selection. In this sense, the complete KA for updating the DEA-ontology associated with the OLeKSS is summarized in the following three meta-processes: 1) The previously attained ontology was used as an input for another SMOL process from a previously developed ontology that belongs to a similar domain. This ontology -LUBM- was retrieved via Internet (Swoogle browser). A new version of DEA-Ontology was reached by ontology matching [22]. 2) The DEA-Ontology updated in point 2, was further enriched through an additional SMOL process from a corpus of documents with 480 Educational Journal articles selected by expert-users [24]. This nal DEA-Ontology version is used as a host for the case study. 3) Finally in this case, the RDB named RDB-IUTEPAS has been used as a KSO. RDB-tables have been selected from this RDB, which has been developed for a real academic information system that currently operates in a small-scale university institution, identi ed by its acronym as IUTEPAS which is established in Cagua in Venezuela. At present, it has about 1000 students and 110 professors [30]. We most point out that we use of database as sources improve the knowledge updating process due to the use the information not only from the database schema but also from the database content. SMOL from RDB was applied to this case study throughout seven phases, including three decision points. The Phase VIII (OLeKSS con guration) will be applied in a future case. However, a special emphasis has been given to SMOLPhase I about methodology strategy drafting and selection. Indeed, a set of 12 RDB-IUTEPAS tables related with the professor subdomain has been chosen. The RDB models obtained have technical compatibilities -links- with the two selected conversion-tools cited below. Particularly, the MRs were selected in this Phases I, including tools to be applied. Respectively, in the Table 2 under the [46] criteria, we have considered
11 also some details about the RDB's Conversion-tools reviewed and tested, according some learning levels which could be reached through them (not included by space limitation). Finally, RDBToOnto and ODEMapster tools were selected. This methodological strategy selection was conceived from a combination of bottom-up (inductive) learning discovery for the rst ow cycle cited, with a top-down (deductive) of learning recovery from RDB during the second cycle. In the rst cycle of SMOL application, the RDB-IUTEPAS and RDBToOnto tool have been used to discover, to recover and to compare -matching- semantic entities that are found in the RDB-IUTEPAS. For instance, a ontology-subclass (university location and name) about where the professors earn their grades was obtained by data-mining processes. In the second cycle, using the lessons learned in the previous one about relevant Classes, and properties from this RDB, the ODEMapster tool has been applied for re ning semantic correlations between the RDB and the host-ontology. Therefore, ODEMapster helps to learn more about the previous (sub)concepts, converted taxonomies found, and to establish better correlation between the tables-attributes of RDB-IUTEPAS with those concept property equivalents in the host-ontology that is undergoing restructuring. A complete and validated ontology was obtained, with important semantic results improving the quality of query (questions) demanded by end-users (students and Carree's advisers). The lesson learned about OLeKSS Processes applied to the case study shows several relevant aspects: a) there are dierent possible MRs that might be combined during OL process; 2) phase's activities can be registered (log- le) and be accessible support users in reuse and in future related decisions; 3) the RDB used can be registered and grouped to support future and better decisions, associated with querying and updating; and, 4) the data source pre-processing from the RDB always involves a prior user's familiarization, with speci c RDBM tools and the RDB conversion-tools used.
7
Conclusions
In this paper, a new framework about the KSS knowledge social context and OLeKSS components are suggested to support the possibility of KSS improvement by OL. Indeed, OL is considered as an useful way to enhance this required updating process. This proposal satis es the continuous knowledge user's need in a more reasonable, adaptable and e cacious way than the traditional KBSs. Our systemic proposal about e-social OL perspective, heterogeneity of e-KSO for OL, and diversity of MRs to support the KA processes as a whole (OLeKSS), is more integrated and exible in comparison with other structured approaches. We have focused on the database as e-KSO in this paper because this kind of sources are relevant to maintain the knowledge updated using not only the databases schemes but also the database content. In this sense, the SMOL methodology was applied as a systemic option for dealing with all the OLeKSS components. It enables OLeKSS Users to discover, to recover, and to manage the potential knowledge from RDB through OL, to maintain corresponding OLeKSS Products -ontologies- updated.
Bibliography
[1] Abdullah, R., M. Selamat, A. Jaafar, S. Abdullah, and S. Sura: 2008, `An Empirical Study of Knowledge Management System Implementation in Public Higher Learning Institution'. Int. J. Computer Science and Network Security
8, 281290.
[2] Abecker, A., A. Bernardi, H. Maus, M. Sintek, and C. Wenzel: 2000, `Information supply for business processes: coupling work ow with document analysis and information retrieval'. Knowledge-Based Systems
13, 271284.
[3] Agostini, A., R. Bosellim, F. De-Paoli, and R. Dondi: 2005, `Situation-based Knowledge Presentation for Mobile Workers'. In: Proced. World Academy of Science, Engineering and Technology, Vol. 3 January.
[4] Antony, S. and R. Santhanam: 2007, `Could the use of a knowledge-based
43, 141151.
system lead to implicit learning?'. Decision Support Systems
[5] Astrova, I., N. Korda, and A. Kalja: 2007, `Rule-Based Transformation of SQL Relational Databases to OWL Ontologies'.
Proceedings of the 2nd
International Conference on Metadata and Semantics Research. pp. 116.
[6] Aurum, A., F. Daneshgar, and J. Ward: 2008, `Investigating Knowledge Management practices in software development organisations - An Australian experience'. Information and Software Technology
50, 511533.
[7] Blackman, D. and M. Kennedy: 2009, `Knowledge management and eective university governance'. Journal Of Knowledge Management
13(6), 547563.
[8] Buitelaar, P. and P. Cimiano: 2008, Ontology Learning and Population: Bridging the Gap Between Text And Knowledge. IOS Press, Netherland.
[9] Callaos, N.: 1992, `A Systemic System Methodology'.
In: International
Conference on System Research Informatic and Cybernetics, Baden-Baden, Germany.
[10] Callaos, N. and B. Callaos: 2006, `Designing with a System Total Quality'.
In: Int. Conference on Information System Analysis and Synthesis,
(ISAS'06), Orlando-USA. pp. 1523.
[11] Cerbah,
F.:
2008,
`Mining
the
Content
of
Learn Ontologies with Deeper Taxonomies'.
Relational
Databases
Proceedings
of
the
to
2008
IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent
1, 553557.
[12] Cimiano, P.: 2006, Ontology Learning and Population from Text: Algorithms, Evaluation and Applications. Springer-Verlag New York, LLC.
[13] Dahlem,
N.
and
A.
Hahn:
Methodologies-A Survey'.
2009,
`User-Friendly
Ontology
Creation
15th Amer. Conf. on Information Systems,
California-USA pp. 19.
[14] Davies, J., A. Duke, and Y. Sure: 2004, `OntoShare: An Ontology-based Knowledge Sharing System for Virtual Communities of Practice'. Journal of Universal Computer Science
10, 3, 262283.
[15] Ehrig, M.: 2007, Ontology Alignment: Bridging the Semantic Gap. SpringerVerlag.
13 [16] Euzenat, J., A. Mocan, and F. Share: 2007, Ontology Management, Ontology Alignments: An Ontology Management Perspective. Springer.
[17] Feldkamp, D., K. Hinkelmann, and B. Thonssen: 2010, The Modelling of Knowledge-Intensive Processes Using Semantics. Book. Springer-Verlag.
[18] Finquelievich, S. and A. Prince: 2006, `Las universidades argentinas en la sociedad del conocimiento'. Technical report, Telefonica, Buenos Aires, Argentina. [19] Gacitua, R. and P. Sawyer: 2008, `Ensemble Methods for Ontology Learning - An Empirical Experiment to Evaluate Combinations of Concept Acquisition Techniques'.
In: Computer and Information Science, ICIS 08. 7th
IEEE/ACIS Int. Conf.
[20] Gaines, B.: 1990, `Knowledge-support Systems'.
Know.-Based Syst.
3-4,
192203. [21] Gil, R.: 2009, `New Systemic Methodology Framework for Ontology Learning (in Spanish)'.
Master's thesis, Dept.Computer Science, University of
Granada, Spain. [22] Gil, R., A.Borges, L. Ramos, and L. Contreras: 2008, `Ontologies Integration for University Institutions: Approach to an Alignment Evaluation'. In: 19th Australian Conference on Software Engineering.ASWEC'08. pp. 570578.
[23] Gil, R., A. Borges, and L. Contreras: 2007, `Shared Ontologies to Increase Systems Interoperatibiliy in University Institutions.'. In: IMCSIT, International Multiconference On Computer Science And Information Technology, Polonia.
[24] Gil, R., L. Contreras, and M. Martín-Bautista: 2010a, `A Systemic Methodology For Ontology Learning: An Academic Case Study And Evaluation'. IC3K, KEOD-2010, Oct. 25-28, Valencia-Spain
1, 206212.
[25] Gil, R., M. Martín-Bautista, and L. Contreras: 2010b, `Applying An Ontology Learning Methodology To A Relational Database: University Case Study'. 4th IEEE Int. Conf. (IEEE-ICSC-2010), Pittsburgh-USA. [26] Gliozzo, A., M. S. C. Caracciolo, M. D-Aquin, W. Peter, J. Voelker, M. Dzbor, E. Mota, A. Gomez-Perez, P. Haase, W. Water eld, J. Contreras, M. Grobelink, J. Euzenat, H. Cunning, S. Staab, A. Gangemi, J. Angele, M. Iglesias, T. Lobo, and A. Lopez: 2007, `Results from Experiments in Ontology Learning including Evaluation and Recommendation'. Technical report, NeOn-Project.org. [27] Gómez-Pérez, A. and D. Manzano-Macho: 2005, `An overview of Methods and Tools for Ontology Learning from Text'. Knowledge Engineer. Rev.
19,
187212. [28] Haase, P., J. Volker, and Y. Sure: 2005, `Management of dynamic knowledge'. Journal Of Knowledge Management
9, 97107.
[29] Holz, H., O. Rostanin, A. Dengel, T. Suzuki, K. Maeda, and K. Kanasaki: 2006, `Task-based Process Know-how Reuse and Proactive Information Delivery in TaskNavigator'. CIKM'06, November 5-11, 2006, Arlington, Virginia, USA.
1, 522531.
[30] IUTEPAS: 1996, Instituto Universitario de Tecnología Pascal.
Insti-
tuto Universitario de Tecnología Pascal, Acta Constitutiva, Web pages http://www.iutepas.8k.com/.
14 [31] Jeusfeldand, M., M. Jarke, and J. Mylopoulos: 2009, Metamodeling for Method Engineering. Cambridge (USA): The MIT Press.
[32] Jung, S., J. Kim, and J. Herlocker: 2004, `Applying Collaborative Filtering for E cient Document Search'. IEEE/WIC/ACM Int. Conf. on Web Intelligence.
[33] Lee, H., J. Kim, and J. Koh: 2009, `A contingent approach on knowledge portal design for RD teams: Relative importance of knowledge portal functionalities'. Expert Systems with Applications
36, 36623670.
[34] Liu, D. and I. Wu: 2008, `Collaborative relevance assessment for task-based knowledge support'. Decision Support Systems 4
44, 524543.
[35] Liu, D., I. Wu, and K. Yang: 2005, `Task-based K-Support System: disseminating and Sharing task-relevant knowledge'. Expert System with Applicat.
29, 408423.
[36] Middleton, S., NR., Shadbolt, and D. De-Roure: 2004, `Ontological user pro ling in recommender systems'. ACM Transactions on Information Systems
21, 5488.
[37] Music, D., J. Azemovic, and M. El-Zayat: 2009, `Components of the e cient eUniversity system'. 2nd IEEE International Conference on Computer Science and Information Technology
1, 3236.
[38] Navigli, R. and P. Velardi: 2006, `Ontology Enrichment Through Automatic Semantic Annotation of On-Line Glossaries'. Science
4248/2006, 126140.
Lecture Notes in Computer
[39] Nonaka, I. and H. Takeuchi: 1995, The Knowledge Creating Company: How Japanese Companies Create the Dynamic of Innovation.
Oxford Univ.
Press. [40] Novacek, V., L. Laera, and S.Handschuh: 2007, `Semi-automatic Integration of Learned Ontologies into a Collaborative Framework'. In: Proceedings of IWOD/ESWC 2007, in conjunction with ESWC 2007.
[41] Nováèek, V., L. Laerab, S. Handschuha, and B. Davis: 2008, `Infrastructure for dynamic knowledge integration-Automated biomedical ontology extension using textual resources'. Journal of Biomedical Informatics
41,
816
828. [42] Noy, N. and M. Klein: 2004, `Ontology Evolution: Not the Same as Schema Evolution'. Knowledge and Information Systems
6, 428440.
[43] Phillips-Wren, G., M. Mora, G. Forgionne, and J. Gupta: 2009, `An integrative evaluation framework for intelligent decision support systems'. European Journal of Operational Research
195, 642652.
[44] Ramos, L. and R. Gil: 2010, `Information System to Support the Management of the Distance Education'. Télématique; Revista Electrónica de Telemática
9, 98119.
[45] Royero, J.: 2007, `From e-Government to e-Universities: a South American viewpoint'. Revista de Universidad y Sociedad del Conocimiento
4(2), 115.
[46] Sahoo, S., W. Halb, S. Hellmann, I. Kingsley, T. Thibodeau, S. Auer, J. Sequeda, and A. Ezzat: 2009, `A Survey Of Current Approaches For Mapping Of Relational Databases For RDF'. W3C RDB2RDF incubator Group.
15 [47] Salazar-Torres, G., E. Colombo, F. C. D. Silva, C. Noriega, and S. Bandini: 2008, `Design issues for knowledge artifacts'. Knowledge-Based Systems
21,
856867. [48] Sallis, E. and G. Jones: 2002, Knowledge management in education: enhancing learning and education. Kogan Page Lim.UK & Stylus Publishing,USA.
[49] Simperl, E., C. Tempich, and D. Vrandecic: 2008, A Methodology for Ontology Learning Chapter of Book Ontology Learning and Population.
IOS
Press, Buitelaar and Cimiano Eds. [50] UNA: 2005, `Plan Estrategico de La Universidad Nacional Abierta:Presente y Futuro desde un Punto de Vista Colectivo'. Technical report, Universidad Nacional Abierta-UNA: Gaceta Universitaria Venezuela. [51] Von-Krogh, G.: 2009, `Individualist and collectivist perspectives on knowledge in organizations: Implications for information systems research'. Journal of Strategic Information Systems
18, 119129.
[52] Zahrawi, A. and Y. Yahya: 2009, `A Framework for Knowledge Management System in Higher Learning Institution: A Case Study of National'. International Conference on Electrical Engineering and Informatics pp. 325328.
1 [53] Zhen, L., Q. Huang, and Z. Jiang: 2009, `Recommender system based on 1
work ow'. Decision Support Systems
48, 237245.
1 1
Fig. 4.
OLeKSS model is represented in UML by Class Diagram
