ONTOLOGIES IN KNOWLEDGE OFFICE SYSTEMS

Ivan Polášek

and Jozef Kelemen

2, 3

Institute of Informatics and Software Engineering, Slovak Technical University, Bratislava, Slovakia

VSM College of Management, Bratislava, Slovakia

Institute of Computer Science, Silesian University, Opava, Czech Republic

Keywords: Knowledge-based systems, Ontology, Knowledge management, Knowledge managing systems.

Abstract: The paper shows one concrete implementation of Knowledge office, where the control of the company and

its knowledge are implemented within the real user-defined document workflow and content management

system. The new planned module is designed to extract the existing knowledge from various types of

documents (source code, documentation, agreements, etc.) by using prepared domain and document

ontology and the Knowledge System to help the user to create new document.

1 INTRODUCTION

The rapid progress in developing and application of

knowledge-based systems in different areas of the

research, industry and administration caused several

problems concerning the management of knowledge

(Kelemen and Hvorecky, 2008).

Knowledge management (KM) means the large

spectrum of activities (McElroy, 2003) connected

with management of company’s shared knowledge

(decomposition, distribution, innovation, acquisition,

accessibility, preservation, etc). The direct use of

knowledge is often shifted to information

technology – to the knowledge-based systems

(Stefik, 1995). For the future, knowledge managing

systems (KM Systems, in short) seem to be a

promising field of research and engineering, and

then perhaps also of huge applications of developed

systems.

2 KM SYSTEM – A CASE STUDY

Gratex Knowledge Office (GKO) is designed for the

management of the company and its knowledge

base, control of the company’s internal processes

and project management. It can be used in a wide

variety of companies with diverse specializations.

GKO.NET is now a user-defined document

workflow and a content management system. Its

variability and document distribution make the

company’s control system more effective and

transparent. It enables internal information sharing

and management, based on predefined unified

procedures and regulations. Information is

transmitted via electronic documents of diverse

types and saved on the central server. These types of

documents register the development of internal

processes in companies. The system enables

definition of roles and powers of employees

transparently. It is an appropriate tool for global

organization control and effective teamwork.

GKO.NET offers opportunities for effective

management of key company processes, such as:

• Standard processes (such as business

administration, decision-making processes,

quality management, purchasing, sales, etc.).

• Safety management (management of

information systems and storages, assignment

of system rights and accesses to individuals

and groups, risk monitoring, monitoring of

weaknesses, threats, and damages, etc.).

• Document administration (registering,

sharing, updating, backup and printing of

various documents, and templates

administration).

• Human resources (hiring, profiles,

qualification, trainings, and courses etc.,

availability information, such as attendance,

absences, sick leave, business trips, payroll

administration, etc.)

• Asset management (registration and

categorization of assets, allocation of work

tools to the employees).

400

Polášek I. and Kelemen J. (2009).

ONTOLOGIES IN KNOWLEDGE OFFICE SYSTEMS.

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development, pages 400-403

DOI: 10.5220/0002293704000403

 SciTePress

Mail Client

GKO Web

Service

GKO WinForm

Admin

GKO WinForm

Client

GKO Reminder

GKO Workflow

Engine

GKO Doc

Reader

SMTP Mail

Server

GKO Store

Engine

CodeBase

ReminderDB

Other

RDBMS

SQL

Server

Index Server

COM+

SMTP

Mail notification

SOAP

Standalone

AI system

Knowledge

Miner

Kohonen

Neural

Networks

Ontology

Software

Special

Searching

Engine

Figure 1: The schematic view of the GKO.NET.

• Project management (definitions of project

teams, scopes of delivery, deadlines, risks,

assumptions, documentation, task planning,

recording suggestions and changes, quality

management, controlling etc.).

Access of users to GKO.NET is simple, enabled

through a common network, or the Internet, and its

implementation requires no significant changes in

the existing infrastructure. The application is easy to

adapt to specific customer needs. It provides for a

flexible administration and specification of security

rights and rules. It can be integrated with other

systems. The overall scheme of the GKO.NET in the

context of some other support systems is depicted in

Figure 1. Standalone AI system communicates with

business layer, which contains business logic (Doc

Reader, Workflow Engine, Store Engine, Reminder).

3 HOW GRATEX KNOWLEDGE

OFFICE WORKS

At the beginning of the integration, it is important to

identify needful entities (Project, Request,

Document, Task) with their state spaces (Initial,

Canceled, Completed, Frozen, Rejected, Quality

Assurance Approved, Frozen, In Progress) and

report criteria, data joins and user filters. This is the

way to create the model of the company and monitor

its life.

For all users, we need to create individual

entities with attributes and their State Spaces

(workflows) with the special activities in the states.

This is the first step to implement precise knowledge

management of the individual company.

The next step is to implement input and output

forms with their rules, triggers and criteria to report

the actual situation in the company.

They are clustered primarily as the project groups,

then secondarily by their states and date, but they

could be ordered in another way.

For example, we prepare for a client the program

Support Server with the objects Client, Incident and

Expert with their attributes and state spaces. Then

we can generate reports to monitor the support

processes with the various states (In Progress,

Solved), and criteria values (Client = IDClientID or

Name = <ClientName>, Importance =

BusinessDown), and aggregate them by various

ONTOLOGIES IN KNOWLEDGE OFFICE SYSTEMS

401

Domain &

Document

Ontologies

Extracted

Knowledge

*.*

*.doc

*.pdf

New

document

Extracting

Knowledge

Searching

Relevant

Knowledge

Creating

New

Document

Source Code

(Java, C#, ...)

Analytical,

Technical or

User

Documentation

Tasks & Plans

Agreements

Knowledge Systems

Figure 2: The process of extracting knowledge and creating new document.

Figure 3: Semantic model as a draft for domain/document ontology.

states or clusters.

The new planned module is supposed to manage

the existing knowledge in documents in the

institution in order to help the user to create new

ones. Figure 2 shows the process of extracting

knowledge from various types of documents by

using prepared ontologies and the Knowledge

System (in the left side of the schema), and the

process of creating new document and searching

relevant knowledge from prepared knowledge base

KEOD 2009 - International Conference on Knowledge Engineering and Ontology Development

402

in the right part of the diagram. Process Searching

Relevant Knowledge offers interesting information

and knowledge as the parts of other documents for

creating the new one. It needs extracted and indexed

knowledge from existing documents, parsed by the

ontology, and the knowledge software. The New

Document item could be documentation, agreements

or source code.

Figure 3 shows the first draft of the semantic

model as a basis for the domain and the document

ontology. It is familiar with the other systems in this

product line (Assali, 2007) in some analogical

features.

We can use the relationship between the Project,

Document and the Keyword to prepare knowledge

mining with the keyword vectors using graph and

clustering algorithms, the Kohonen Self Organizing

Maps (using keyword vectors) and the mining of

associative rules to find relevant documents.

Also we can find the relation between the users,

authors and the documents to prepare corresponding

rules for the user history and authors of different

sources for the knowledge system. Except for the

title and the authors, document contains also the

modules (chapters, appendices, classes, packages,

paragraphs). Concrete structure is specialized in the

independent types of the document with their proper

ontology (the source code is quite different from the

analytical document or agreement).

In this manner, these dynamic (mainly the new,

created one) document ontologies increase the

power of the system to deal with the given

document, how to understand its content and its

relevance to other documents in the systems

depository. Domain ontology for the whole system

could be also dynamic, to map the various type of

the companies and areas (economical processes,

SLA, financial institutions, law companies,

manufacturing corporations, construction

companies, software houses, etc.).

The visual model of ontology is creating for our

designers, using model close to object-ontology

mapping (Bartalos and Bielikova, 2007), but we

need the code for the parser in the next step.

4 CONCLUSIONS

At the present time, Gratex Knowledge Oficce is

working 24 hours per day and supports the

management of five diverse companies (Arca-

capital, Hornex, Milking, Elvea, Gratex International

(all internal and economic processes) and Gratex

SLA (Service level agreement for Allianz)). In GKO

we implement the knowledge management to the

system using DMS and the workflow development.

Our new designed module would help to use

knowledge from the documents to create the new

one. First pilot was created in cooperation with

Slovak Technical University and Gratex

International, next version will be published with the

closed set of ontologies for the software company

domain, type of documents and problem areas

(banking, insurance, etc.).

ACKNOWLEDGEMENTS

The authors’ research on the subject of this

contribution is partially supported by the Scientific

Grant Agency of the Slovak Republic grant No.

VG1/0508/09.

REFERENCES

Assali, A., Lenne, D., Debray B., 2007. KoMIS: An

ontology-Based Knowledge Management System for

Industrial Safety, In 18th International Conference on

Database and Expert Systems Applications (DEXA

2007), Regensburg, pp. 475-479.

Bartalos, P., Bielikova, M., 2007. An Approach to Object-

ontology Mapping. In Software Engineering in

Progress, CET-SET 2007, Poznan.

Kelemen, J., Hvorecký, J., 2008. On knowledge,

knowledge systems, and knowledge management. In

Proc.9

International Conference. on Computational

Intelligence and Informatics, CINTI 2008, Budapest

Tech, Budapest, pp. 27-35.

McElroy, M. W., 2003. The New Knowledge

Management. Elsevier, Amsterdam.

Stefik, M., 1995. Introduction to Knowledge Systems.

Morgan Kaufmann, San Francisco, Cal.

ONTOLOGIES IN KNOWLEDGE OFFICE SYSTEMS

403