THE RESEARCH OF KNOWLEDGE MANAGEMENT SYSTEM
FOR SUPPORTING HYPER TEXT TYPE OF ENTERPRISE
ORGANIZATION
Masaru Iguchi
1
and Yoshiyuki Kotani
2
1
Department of Computer and Information Sciences, Tokyo University of Agriculture and Technology, Tokyo, Japan
2
Institute of Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan
Keywords: Knowledge Management, Document, Hypertext, Organization, Ontology.
Abstract: In recent years, many enterprises have used knowledge in their management. Due to globalization, the
organizational structure of enterprises is trending towards integration of the conventional bureaucracy type
with a “Hyper Text type”, in which the organization is formed for each project. Such structural change
brings a need to also change the knowledge management of enterprises. This paper defines enterprise
knowledge as documents, and proposes a document search system which suits the above “Hyper Text type”
organization.
1 INTRODUCTION
Many enterprises consider the use of knowledge in
their management to be increasingly important. On
this point, the Organization for Economic Co-
operation and Development (OECD) claims in its
report “The Knowledge-based Economy” that
“Knowledge is now recognized as the driver of
productivity and economic growth.” This clearly
indicates the importance of knowledge in economic
activities (OECD 1996). Enterprises which develop
and mass-produce industrial products especially
need to invest a large amount of money in research
and development, and consider knowledge as an
asset. On the other hand, in activities of such
enterprises, the time interval from launching a new
product to the next has been shortened. For a project
to develop a new product according to the shorter
time interval, one or two members are selected from
each division (development, purchasing, production,
sales, etc.) to form a taskforce. While the taskforce
members belong to their own divisions, they work
for the project until it is completed. Nonaka and
Takeuchi considered this organizational structure a
“Hyper Text type” organization, and took Japanese
enterprises which aimed to integrate bureaucracy
with taskforces as examples for their study (Nonaka
and Takeuchi, 1995). This paper defines knowledge
as documents that express knowledge, and discusses
a system that supports document management for
the enterprises which have the above “Hyper Text
type” organization. First, relevant studies are
presented in Chapter 2. The outline of the document
management system in the “Hyper Text type”
organization and an overview of the proposed
document management system are presented in
Chapter 3. The conclusion of this paper and future
study themes are described in Chapter 4.
2 RELATED WORK
2.1 Knowledge and Enterprise
Organization
Accumulating as knowledge the developmental
abilities which demonstrate the strength of
enterprises has been increasingly important for the
competition and success of enterprises.
Consequently, improvement to the traditional,
organizational structure, namely bureaucracy, is
considered necessary. However, instead of
eliminating the bureaucracy, aiming for an
organizational structure that is mutually
complementary with taskforces is considered
necessary. Such an organizational structure is the
312
Iguchi M. and Kotani Y..
THE RESEARCH OF KNOWLEDGE MANAGEMENT SYSTEM FOR SUPPORTING HYPER TEXT TYPE OF ENTERPRISE ORGANIZATION.
DOI: 10.5220/0003635403120317
In Proceedings of the International Conference on Knowledge Management and Information Sharing (KMIS-2011), pages 312-317
ISBN: 978-989-8425-81-2
Copyright
c
 2011 SCITEPRESS (Science and Technology Publications, Lda.)

“Hyper Text type” organization, which refers to
“Hyper Text” originally developed in the computer
science field (Nonaka and Takeuchi, 1995).
Figure 1: The hypertext organization (Nonaka and
Takeuchi, 1995).
Figure 1 shows the “Hyper Text type”
organization. The layer in the center of the figure is
a business system layer in which normal routine
work is done. As the bureaucratic structure is
suitable for its efficiency, this layer is a hierarchical
structure. The top of this structure is a project team
layer, in which many project teams are engaged in
product development and the like. Members of these
teams are gathered from many divisions of the
business system layer, and engage in one project
until it is finished. When the team finishes the task,
it returns to the knowledge base layer, draws up a
knowledge inventory describing successes and
failures of the project, and returns to the business
layer. Based on his work experience at an enterprise
where he worked, the writer thinks that even if the
aforementioned knowledge is prepared, if the
situation does not enable its reuse, in cases where
part of the description needs revision or where
another project team wants to refer to it, accessing
the inventory is very difficult. Mainly using cases of
European enterprises, Enkel and others proposed a
knowledge network for the growth of enterprises
consisting of three layers (Enkel, 2007). Those cases
present a facilitating condition summarizing a
management methodology practiced by enterprise
management, a knowledge process regarding
knowledge work by interaction among employees,
and a reward structure and meeting methods and
communication methods serving as tools to support
the knowledge process. The framework designed by
Enkel and others is a model to relate and
comprehensively manage the business strategy with
systems and tools that support the business process.
However, this framework is different from an
enterprise organizational structure which is the target
of the system proposed in this paper. In addition to
this, Duncan (Duncan, 1979) and Ackoff (Ackoff,
1989) advocate building relationships among data
for intra-enterprise knowledge distribution, but both
of these differ form the proposal of this paper in the
intended organizational structure and the data link
structure.
2.2 Knowledge Management System
In recent years, computerized documents and email
have been used for in-house communication. In
enterprises, in-house knowledge distribution based
on a community called a “Community of Practice”
(CoP) or “Community of Interest” (CoI) proposed
by Wenger and others has been introduced into
business management (Wenger, 2002). Here, CoP
signifies a community where knowledge effective
for business performance is exchanged, and CoI
signifies an informal community where knowledge
of wide interest to employees is distributed. Based
on this principle, Lavoué proposed a model that
connects produced knowledge to activities proposed
by the CoP (Lavoué, 2009). On the other hand,
Benjamins and others used techniques of ontology
engineering in their approach to knowledge
management systems (Benjamins, et al, 1998). The
approach allows the discovery of knowledge that is
not explicitly known but can be deduced based on
general knowledge which is captured in the
ontology). Figure 2 shows overview of the approach.
Figure 2: An Ontology-based approach (Benjamins, 1998).
An ontology of the subject matter must build,
which is used to characterize the subject matter, for
instance to fill the ontology with instances. A
intelligent web crawler receives a query in term of
the ontology, consults them using the ontology and
generates an answer. Though the techniques of
ontology engineering are applied to the proposed
document management system described later in this
THE RESEARCH OF KNOWLEDGE MANAGEMENT SYSTEM FOR SUPPORTING HYPER TEXT TYPE OF
ENTERPRISE ORGANIZATION
313

paper, the composition of document management is
considered one of the ontologies in this paper. Thus
this paper differs in that it does not newly define
domain ontology, and in that a document itself is
considered an instance.
3 OUTLINE OF THE PROPOSED
DOCUMENT MANAGEMENT
SYSTEM
In recent years, as a means to certify the quality
level of business administration activities, an
increasing number of corporations have been
obtaining ISO9001 certification (ISO, 2009).
ISO9001 contains guidelines for processes of
business activities and detailed management
activities. Each corporation formulates its own in-
house work process and management system, based
on these guidelines. Document control is assessed as
a part of the management system. For example,
regarding documents such as information prepared
from the business activities, an enterprise manager
must stipulate the preparer of the document, the
approver, and assign a management number to
stipulate the storage system and time period. As for
the concept of the document management system
proposed in this paper, the target of knowledge
obtained from the project activities in the
organizational structure of the aforementioned
“Hyper Text type” are documents under the control
stipulated in ISO9001 (hereinafter, “Controlled
Document”).
It is considered that the Controlled Document
should enable search to meet needs of people in
various positions such as business managers and
staff, and yet it should easily be reused. As for the
document management system, the management
system of a real European-financed Japanese
corporate for which the author works is used as an
actual case for the development of the argument.
3.1 Structure of Document
Management
The document management structure covered by the
system proposed in this paper is illustrated in
Figures 3 and 4. Figure 3 outlines processes from
metadata to the Controlled Document. For instance,
the author at the R&D division shown in Figure 4
drafts a document in the form of metadata by
electronic means such as MS Word. The author, if he
or she belongs to an R&D division, assigns a
document number of the division as a filename (e.g.
TNA2011A001) stipulated by the corporation based
on ISO9001, and files the document name as
“Document A” (hereinafter, “Doc. A”) in the
Document Server (Figure 4). In general, each
corporation prepares the Controlled Document
according to the request of a milestone in work
processes (e.g. a meeting for reporting the progress
status or settlement of a project). This should be the
formal document stipulated in the corporation’s
ISO9001. It should also be accessible to all
pertaining employees, as well as reusable. Figure 4
shows that the Hyper Text type “organization”
described in the previous chapter is converted into
“document” as an example of a certain corporation.
Figure 3: Example: Process of publishing a document.
The document is described as the work process and
document issued by respective divisions that utilize
the process, and each of them is prepared in the
same process as that shown in Figure 3. Here, the
relationships between the Hyper Text type
corporation responsible for a certain project A, the
“Project Team” and the document required in the
team is illustrated in Figure 4.
Figure 4: Document Management Structure.
With the aim of completing a certain task called
project A, the Project Team consists of one or two
representatives (non-full-time in general) dispatched
KMIS 2011 - International Conference on Knowledge Management and Information Sharing
314

from R&D, Purchasing, Production and Marketing
divisions respectively. This task acts based on the
work process stipulated by this corporation.
Therefore, the leader of this project must prepare a
briefing paper for meetings, during which the
progress status is reported based on each milestone
of the work process. Consequently, this briefing
paper is important knowledge for the corporation,
with management information such as specifications
and investment necessary for the completion of the
project A, technical issues, methods to solve such
issues, and the results. The briefing paper can be
called a combination of documents prepared by
respective team members. However, as the team
members who prepare respective documents are
dispatched from other divisions, they will assign
document numbers stipulated in document
management of respective divisions, and file the
documents in folders of respective divisions of the
document server. As the Hyper Text type
organization (in this paper, the “project team”) is
dissolved at the completion of the task, it is very
difficult to retrieve the combination of documents
prepared in the organization, even if it is filed in a
dedicated folder and is reviewed fully at that time, if
a person responsible for the maintenance including
revision after the dissolution is unknown, or if
leaders of other projects want to refer it, or heads of
respective divisions want to refer it.
3.2 Outline of Proposed Document
Management System
The biggest feature of this system, the ontology that
defines the feelings of the search, by matching as a
set of documents defined in the domain is to increase
the efficiency of the search. As described above, the
basic model for this system is the Japanese company
which the authors belong to, so documents handled
are written in the Japanese language. Also, the needs
described above were confirmed by questionnaires
from the company’s employees (100 people) at
various levels. It was confirmed that those needs can
be categorized into the following three broad groups.
a. The document number register is linked to each
document with a number assigned thereby.
b. Documents created by a project team can be
searched in a set.
c. Documents created by a project team and other
documents separately created can all be searched
by natural language.
For group an above, the matter can be solved by
relating documents to the document number register
in simple 1:1 links, so this paper’s research covers
groups b and c, as discussed below. Its system
configuration is shown in Figure 5. “GUI Search and
Answer” are items to select when a searcher
searches for a document. For a real search engine, it
builds “Search Engine by Ontology” for adaptation.
The reason to use ontology in the search engine is
considered to be that for the purpose of obtaining
their respective ISO9001 certifications, the business
organization discussed in the present paper to be
linked in the search engine and the document
management system are classified and formulated by
a proper tree structure, and the action of document
search itself is what should be clarified by ontology.
“Search Engine by Ontology” clarifies the
searcher’s search goals, and according to those
goals, searches for needed documents from the
“Document Servers” which stored the documents,
thereby replying to the searcher. Figure 6 shows the
Class structure outline of “Search Engine by
Ontology”.
Figure 5: Outline of proposed Document Management
System.
Figure 6: Outline of Class structure outline of “Search
Engine by Ontology
.
“Document Search” is under the highest conceptual
class is “Search”. The conceptual classes directly
under this were obtained from results of the
questionnaire mentioned above, so they are
THE RESEARCH OF KNOWLEDGE MANAGEMENT SYSTEM FOR SUPPORTING HYPER TEXT TYPE OF
ENTERPRISE ORGANIZATION
315

comprised as Upper and Lower concept. “Single
Document” does a search using the registration
number register which has 1:1 links with documents.
“Multi Documents” is comprised of a group of
multiple independent documents. For example,
specific documents are needed in regular meetings
concerning decision making during a product’s
development, so these define the concept of
“Meeting”, and the index of documents used in each
meeting is defined in a sub-class. As an instance, it
has a “Table of combination” which specifically
describes that index. That is linked to each
document stored in a “Document Server” described
below, so they are extracted and provided to the
searcher. “Key Word” provides to the searcher
abstract search items which can be specified by the
company’s departments. “Key Word” also provides
the “Status Class” which supports the intentions of
the searcher when narrowing down the search results.
The technology of the key word search is now on
consideration. However, the planned algorithm is
shown in below.
Planned Algorithm
S
G
={Concept
1
,Concept
2
,･･･, Concept
h
}
S
G
is the set of concept in the domain Ontology G.
"Concept
i
"is the concept of the domain Ontology.
T
G
={t
1
,t
2
,t
3
･･･t
n
}
"T
G
"is the set of terminology which are used in the
domain Ontology G and "t
i
" is the term.
P expresses the matrix of relation between the
concept and the terminology.
P
i j=
”1” If the “t
i
” means “Concept
j
”.
= “0” the other.
Calculate “word vector” d
k
by counting the
frequency d
k i
that the term t
i
appears in the
document k which should be searched.
d
k
=
t
(d
k 1
,d
k 2
, ･･･ , d
k n
)
Define the document ontology vector D
k
by matrix
M which transfers from “word area” to “concept
area”.
a (i) =
t
C
i t
P d
k
a (i) is the “Spreading
Activation”.
In case of implementation of the activation to λis
shown in below,
M=(I +
t
C +
t
C
2
+ ･･･ +
t
C
λ
)
t
P
“I” is a unit matrix.
D
k
= M d
k
Then, “λ” should be decided by the depth of the
layer of the domain ontology. When “Search
Ontology” is given as vector Q, the similarity “Sim
(Q, k) “is calculated the same with the “word
vector” d
k.
“Q” is calculated by the same process with the
document ontology vector.
4 CONCLUSIONS AND FUTURE
WORK
The present paper’s proposed document
management system is still under construction, but it
is considered to have the following advantages.
* As the search engine expressed clearly by “Search
Ontology” is used, there is no need to reconstruct
many documents stored in the document server.
* Even if the maintenance of search items is required,
only the classes constructed by the part-of are
modified, so the modification work is easy.
* After the completion of the task, documents used
in the project are reusable in the latest edition, even
if they are revised.
* Clients in divisions not involved in the project can
also refer to documents in respective projects or past
cases.
With the globalization of management and
speeding up of product development, Japanese
corporations have recently shifted their
organizations from conventional bureaucracies to a
flat Hyper Text type, or are working to fuse these
two types. Therefore, the document management
system aimed at in the present paper is considered
useful not only in Japanese corporations, but also in
the world’s business administration in the future.
From this time forward, we will continue to study
the following themes.
* Clarify the document search, and construct the
sub-classes of “Search Ontology”.
KMIS 2011 - International Conference on Knowledge Management and Information Sharing
316

For the clarification of document search, conduct
a questionnaire at the company for which the author
works, analyze the result, and decide on the sub-
classes of the search ontology.
* Break documents into morphemes in the
Document Server, and extract semantic key words.
And start to experiment by planned algorithm with
the above themes, using around 1000 documents,
verify the effectiveness of the document
management system proposed in the present paper.
REFERENCES
OECD, THE KNOWLEDGE-BASED ECONOMY,
Knowledge Based Economy, GENERAL
DISTRIBUTION OECD/GD, Vol. 96, No. 102, 1996.
Ikujiro Nonaka and Hirotaka Takeuchi, The Knowledge-
Creating Company. How Japanese Companies Create
the Dynamics of Innovation, Oxford University Press,
New York, 1995.
Enkel, E., Back, A. and Krogh, G. von, The concept of
knowledge networks for growth, in Back, A., Enkel,
E., Krogh, G. von (eds.), Knowledge Networks for
Business Growth, Springer 2007.
Duncan, R. and Weiss, Organizational Leaning:
Implications for Organizational Design, Research in
Organizational Behavior, Vol.1, pp. 75-13 1979.
Élise Lavoué, E. A Knowledge Management System and
Social Networking Service to connect Communities of
Practice., IC3K 2010, CCIS 128, Fred A. et al. (eds.),
Springer, Heidelberg, p. 310-322, 2011.
Ackoff, R. L, From data to wisdom, J. Appl. Sys. Anal.,
vol. 16, pp. 3-9, 1989.
Wenger, E., McDemott, R. and Snyger, W., Cultivation
community of Practice, Harvard Business School
Press, 2002.
V. Richard Benjamins, Dieter Fensel and Asuncion
Gomez Perez, Knowledge Management through
Ontologies., In proceedings of the Second
International Conference on Practical, Aspects of
Knowledge Management (PAKM), Also in
proceedings of Pacific Asia Conference on
ExpertSystems (PACES'99), LA, CA, 1999.
THE RESEARCH OF KNOWLEDGE MANAGEMENT SYSTEM FOR SUPPORTING HYPER TEXT TYPE OF
ENTERPRISE ORGANIZATION
317