Printer Friendly

A method for enterprise knowledge map construction based on social classification.

INTRODUCTION

With the acceleration of global economy and IT development, enterprises today endeavour to explore better approaches to improve organizational adoption, survival and competence in the new business environment characterized by dynamic, discontinuous and rapid pace of change (Yogesh, 2000). Knowledge management is increasingly viewed as a crucial factor for organizational sustainable competitive advantages (Chase, 1997; Wiig, 1997; Li, 1999a,b; Yogesh, 2000; Yang and Li, 2001; Chen and Li, 2006; Li and Zhao, 2006; Qi et al., 2006; Xu et al., 2006). It represents the recent effort on information management and embodies organizational processes that seek synergistic combination of data and information processing capacity of IT, and the creative and innovative capacity of human beings.

Empirical studies have shown that while organizations learn and create knowledge, they also forget (i.e. do not remember or lose track of the acquired knowledge) (Argote et al., 1990; Darr et al., 1995). Many organizations accumulate knowledge along with their business progresses. Especially the advent of the internet has provided mechanisms for collecting information together, and acquiring knowledge over great distances at any time through new modes of knowledge sharing and enabled partnerships. It is common that enterprises possess a large number of knowledge resources involving complicated structures and the employees have to unfortunately spend so much time and effort on knowledge searching and selection before they can find out what they want. To address these problems, many solutions have been proposed. Knowledge map is one of them.

Knowledge map is an effective knowledge management tool (Davenport and Prusak, 1998; Vail, 1999; Mertens et al., 2003). For the convenience of knowledge navigating and searching, it specifies the captured knowledge and their relationships and displays them in ordered and friendly forms. Currently, the enabling technologies of knowledge map mainly are intranet-based software solutions which combine powerful visualization techniques with database management system (Eppler, 2001). Yet, while the technological implementation can lead to a useful knowledge map artefact, the process of mapping is even more challenging. Essentially, the knowledge mapping process is the process of knowledge organization and classification. Many approaches have been suggested to organize knowledge in organizations that basically fall into two groups. On the one hand, AI methods are suggested to support knowledge modelling and classification, especially for some kinds of web contents. On the other hand, business processes' models are used as a starting point to identify the most critical business knowledge in organizations (Studer et al., 1998). However, in the organizations so far mostly pragmatic approaches are applied. In most of the cases, knowledge classification is determined by a committee in a workshop without much methodical support (Maier, 2004). No matter how the knowledge map is constructed, it has to be characterized with the following factors:

* it can effectively map a large number of knowledge items that are represented in a variety of types of media into some reasonable categories,

* the taxonomy in the enterprise knowledge map has to reflect the characteristics of business process and be comprehensive to knowledge workers.

According to the above requirement, the existing methods are more or less unsuitable for enterprise knowledge map construction. Knowledge mapping is always such a complicated process that we intend to discuss the issue from a systematic perspective. Systems science has been considered as the basis for information systems. A wealth of research in information systems in the framework of systems science has produced an astonishing array of theoretical results and empirical insights, and a large suite of tools and methods (Xu, 2000; Warfield, 2007, 2008). Systems science also promises to be an important foundation of knowledge management. Besides information technologies, human beings are the indispensable component in enterprise knowledge management. They are involved in the knowledge creation, sharing and usage activities. At the same time they interact with each other and learn from each other. It is necessary to put more emphasis on individual's cognition on knowledge map construction. So here, social classification is introduced to assist the knowledge organization of the individual level. Upon that, domain level and inter-domain level will be constructed to show how individual knowledge structure will affect the organizational knowledge structure.

KNOWLEDGE MAP

Many definitions of knowledge maps that we have found in the academic papers are similar, but less specific. Generally, a knowledge map is the display of acquired knowledge and relationships (Davenport and Prusak, 1998; Vail, 1999). The knowledge in knowledge map may involve various shared contents, such as text, graphics, videos, models and data. The relationships among them are determined by linking concepts or topics discovered from these shared contents. Knowledge mapping is defined as the process of associating items of information or knowledge in such a way that the mapping itself also creates additional knowledge (Vail, 1999).

Knowledge mapping strategies differ with respect to the degree of formality that they focus. On the one hand, methods and techniques from the field of AI and computer science are highly formal and represent knowledge in the form of ontologies, domain models or vector space model (VSM) that can be processed by computers. On the other hand, knowledge mapping techniques often primarily serve as a tool for human beings to better understand the structure of important areas of knowledge or competence and their relationships to, e.g. the persons, groups or other organizational units that create, hold, seek, distribute or apply the knowledge. Currently there are three main methods for constructing knowledge maps, including building directories, manually drawing maps and automatic knowledge classification (Ong et al., 2005).

Building Directories

A directory is an alphabetical or classified list of names, addresses and other data. Here the directory denotes the classification of information entities in the form of hierarchy, according to the presumed relationships of the real-world entities that they represent. On the internet a web directory is usually used to organize topics in groups and subgroups, such as YAHOO! and Open Directory Project. It is a simple but effective way to organize a large volume of information, especially when coupled with a search engine. Nevertheless, it is recognized that the interface of directory becomes increasingly difficult for users to navigate as the hierarchy grows larger (Massicotte, 1988; Drabenstott and Weller, 1996).

Manually Drawing Knowledge Maps

Concept maps (Novak and Gowin, 1984) and topic maps (Pepper, 2002) are both like drawings to organize information, in which blocks represent concepts, topics or things and connecting lines represent relationships. Topic map is one of the most important standards (International Organization for Standardization, 2006) for the description of semantics of documents and web resources that foster 'intelligent' information search and processing. The key concepts in topic maps are topics, associations and occurrences, i.e. resources that are linked to topics. These kind of maps can help better organize, display and understand knowledge. However, the manually creation process requires high creator's cognitive skills and significant time and effort. The automatic drawing method is highly desirable.

Automatic Knowledge Classification/Clustering

More recent work constructs the knowledge maps for some kinds of web contents (such as news(Ong, 2005)), employing machine-learning algorithms to cluster the web documents. Document representation and clustering technique are two major issues in text clustering. The VSM is usually adopted to represent documents, where a document is represented as a multidimensional vector, while each dimension corresponds to a unique key term extracted from the documents. A common clustering technique is Self-Organizing Map (SOM), which is an unsupervised neural networks algorithm. Chen et al. (1996) categorized a portion of the internet documents with multilayered SOM to generate a hierarchical knowledge map system. Ong (2005) employed an improved interface combining a 1D alphabetical hierarchical list and a 2D SOM island display to automatically generate a hierarchical NewsMap.

To some extent, the knowledge map construction method based on the automatic text classification is effective. However, when applied in a real organizational environment, it suffers from many problems. First, the method relies much on the linguistic usage. Clearly, in organizational knowledge repository there are many kinds of files, such as drafts, videos, which have few words and consequently are not suitable for this kind of methods. Second, the design of taxonomy may not reflect business needs (Marwick, 2001).

By constructing a knowledge map, it should become possible to examine the knowledge we depend upon on a global scale and from different perspectives (Eppler, 2001). But more important, knowledge map should serve as a navigation of the required knowledge, as it is almost forgotten by most of the designers when they are designing knowledge maps. If we consider the knowledge map construction from the user perspective, the basic idea of constructing a knowledge map of the global enterprise might be old. As an individual employee and a project team is the actual knowledge user and knowledge application context, the knowledge map should put more emphasis on assisting an individual employee, or a team in understanding and using the knowledge available in an organizational setting. It is quite pragmatic to construct a knowledge map from the knowledge application context.

To address the above problems, social classification is introduced into the enterprise knowledge map construction.

INTRODUCING SOCIAL CLASSIFICATION INTO BUSINESS CONTEXT

Social classification, which is also called folksonomy, refers to the collaborative way in which information can be organized on the web. It allows users to publicly add keywords to the shared contents, as it is totally different from the traditional categorizing performed by an authority or authors. Users can not only categorize information for themselves, but also browse the information categorized by others. Today, tagging is a widespread phenomenon popularized by applications such as social bookmarking (Del.icio.us) and social photo sharing (Flickr). Keywords tagging is nothing new; the interesting thing is that when persons tag in a public space, the collection of their keyword/value associations becomes a useful source of data in the aggregate (Gruber, 2005). However, it is discussed in some researches that tagging on the internet has some limitations and weaknesses. For instance, ambiguity can emerge as users apply the same tag in different ways, while the lack of synonym control can lead to different tags being used for the same concept, precluding collocation (Mathes, 2004).

However, tagging is receiving recognition in the intranet applications. The Pennsylvania State University Library has rolled out a social bookmarking service called Penn Tags (http://tags. library.upenn.edu/) for its community. Anyone can browse the items in the database, but only community members can add entries. The Berkman Center for Internet & Society at Harvard Law School launched a social bookmarking site called H20 Playlist (http://h2obeta.law.harvard.edu/). In the corporate world, IBM is developing an enterprise-wide social bookmarking application called 'dogear'. IBM already has a robust internal taxonomy and plans to augment that with folksonomies. It is not a case of a folksonomy replacing a taxonomy. IBM thinks that 'dogear' has the potential to help reveal the interests and expertise of co-workers in order to solve real-world problems, 'dogear' can also help foster communities of practice and increase communication. We argue that social classification would be a fruitful way when it is applied in an organization.

* As the employees, especially those who work in the same or similar domain, share common business goals and have relatively similar business background, the problem in uncontrolled vocabulary will be solved to some extent.

* Social classification generated by employees will facilitate workplace democracy and the distribution of knowledge organization tasks among people actually using them, which will reflect the real situation of knowledge understanding and using from the users.

* Social classification is embedded in the worker's business processes, and could minimize the cognitive load for KM tasks.

* Social classification is most helpful for an enterprise when there is nobody in the 'librarian' role or there are too many unordered contents for few authorities to classify.

* Social classification can provide insights into an individual's expertise and facilitate learning from others.

So here we propose an improved social classification-based method to organize knowledge resources and construct enterprise knowledge map.

OUTLINE OF THE PROPOSED METHOD

In this section, we describe the knowledge map structure and the architecture of the proposed method for knowledge map construction.

Figure 1 shows the proposed knowledge map structure. To illustrate the knowledge and their relationship, our proposed knowledge map is composed of a number of domains, topics, knowledge resources and relationships.

For an enterprise-wide knowledge map, it is very difficult to draw all kinds of knowledge and their categories into one picture. In order to design taxonomy to cover the whole area of interest in enough details, here we define domain as the basic unit to organize knowledge and their relationships. Domain is the context where a specific task, project or business will be carried out and also where knowledge is used, so domain-based knowledge mapping is more meaningful. Similar to topic map, our proposed knowledge map also consists of topics, which represent some concepts in a domain. Topics in different domains are related to each other by associations. A topic may also be related to knowledge resource by its occurrences.

Figure 2 shows the architecture of the knowledge map system. In Figure 2, there are three major tasks associated with the knowledge map construction method: individual knowledge tagging, domain topic selection and inter-domain topic association analysis.

* Individual knowledge tagging is a process of social classification, which is the basis of the enterprise knowledge map construction. When individuals refer to some documents, which are helpful or important for his work, they will tag them. This is a process where knowledge workers organize their knowledge.

* Domain topic selection is a process to construct a knowledge map for a team or department. When the members have classified the knowledge of interest, domain topics will be selected from the tags given by individual through an algorithm.

* Inter-domain topic association analysis is a process to construct the enterprise knowledge map. It is carried out when two topics in different domains are similar with each other to some extent.

[FIGURE 1 OMITTED]

The relevant details will be discussed in the next section.

KNOWLEDGE MAP CONSTRUCTION

In this section, the method to construct the enterprise knowledge map is described in detail. According to the framework mentioned in the last section, the method includes three steps: individual knowledge tagging, domain topic selection and inter-domain topic association analysis.

Individual Knowledge Tagging

Individual tagging is a process for an employee to organize the knowledge of interest and form a personal knowledge map, which records every user's ideas about knowledge classification and his knowledge usage manners. When an employee who works in a specific domain create or experience with a document, he will label the document with one or more proper labels which he thinks could describe the document.

The social classification is formally stated as follows.

[FIGURE 2 OMITTED]

To account for the full environment of social tagging clearly, we define the tagging to be a three-place relation, Tagging (tagger, object, label).

Let Tagging = {[tg.sub.1], [tg.sub.2], ..., [tg.sub.k],..., [tg.sub.s]} be the set of tagging records, where [tg.sub.k] = ([p.sub.l],[d.sub.i], [t.sub.j]), [tg.sub.k](k = 1,2, ... , s) is a piece of tagging record. [p.sub.1] (l = 1,2, ... , h),[d.sub.i](i = 12, ... , m)[t.sub.j](j = 1,2, ... , n) respectively denotes an individual employee, a knowledge item and a tag.

From the individual perspective, all of his tagging records will form a knowledge-tag matrix in Table 1. Here we can regard the tags as attributes of the knowledge items.

In order to describe individual's knowledge classification pattern, we can construct the Hasse Graph of the concept lattice (Sahraoui et al., 1999) and form the individual knowledge map (Figure 3).

Individual tagging is a process to form a personal knowledge map, which records every user's ideas about knowledge classification and

his knowledge usage manners. It means more for the organization. The collection of tagging data will provide great insight into the status of allocating and applying knowledge and help to construct the enterprise knowledge map.

Domain Topic Selection

We have mentioned above that a domain provides a context where knowledge items and their relationships are displayed. When a specific task or project is carried out, a domain appears. As the project is going on, the tag space will consist of many unique terms, words or phrases that are tagged on documents. An example of knowledge tagging records of domain A is shown in Table 2. It is highly desirable to reduce the tag space without sacrificing categorization accuracy.

[FIGURE 3 OMITTED]

According to Yang's study on feature selection techniques (Yang and Pedersen, 1997), information gain has relatively good performance. So here the domain topic is selected by analysing domain members' tagging data and calculating the information gain of each tag. The bigger the information gain of the tag is, the better the tag classifies the document to a category. Assume that the current tag is t, and the set of categories made by domain expert is C = {[c.sub.1], [c.sub.2], ... , [c.sub.m]}, the information gain of tag t is defined to be

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

For the domain A, C is defined by the domain expert as C = {[c.sub.1], [c.sub.2]}, where [c.sub.1] denotes 'important for the domain' and [c.sub.2] denotes 'not important for the domain'. Then the information gain of each tag is calculated as follows:

P([t.sub.1]) = 0.0021, P([t.sub.2]) = 0.0219, P([t.sub.3])

= 0.2983, P([t.sub.4]) = 0.0274, P([t.sub.5]) = 0.0021

So here we select tag [t.sub.3] as the representative attribute for this domain. Of course, it has no problem to select the top three or top five tags.

After selecting the representative tag, we have to link the documents to the topic to show the occurrences. Table 2 can be transformed into a domain knowledge-tag matrix in Table 3.

This matrix is different from Table 1. Let B = {[b.sub.ij]} be the matrix, where

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

u denotes the tagging frequency of tag [t.sub.j] for knowledge item [d.sub.i]. Here, we define [Support.sub.j] to measure the support of tag [t.sub.j] from all knowledge items: Support = {[s.sub.1j], ... , [s.sub.ij], ... [s.sub.mj]}, where

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

So, the support of tag [t.sub.3] is [Support.sub.3] = {1,0.5,0.5,0.75,0.25,0,0,0,0}. Figure 4 is the domain knowledge map.

Inter-Domain Topic Association Analysis

In order to facilitate knowledge reuse among the organization scope, it is necessary to relate the topics of different domains by similarity, which is the process of the enterprise-wide knowledge map construction.

Topic (tag) here is represented in vectors [topic.sub.x] = ([w.sub.1x], [w.sub.2x], ... [w.sub.mx]}, where each [w.sub.ix] is a weight for document [d.sub.i] for topic x and [w.sub.ix] = [s.sub.ix] (i = 1,2, ... , m). Similarity of two topics in different domain equals to cosine of the angle between them, i.e.

sim([topic.sub.x], [topic.sub.y]) = cos ([theta]) = [topic.sub.x] x [topic.sub.y]/[absolute value of ([topic.sub.x]|[topic.sub.y])]

[summation over (i)][w.sub.ix] x [w.sub.ix]/[summation over (i)][w.sup.2.sub.ix][summation over (i)][w.sup.2.sub.iy]

If the similarity between topic x and topic y is greater than a predefined threshold [beta], then we will build a connection between these two topics. Otherwise, there is no association between them. Figure 5 shows the topic association between domain A and domain B (we do not list the tagging data in domain B).

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

KNOWLEDGE MAP SYSTEM IMPLEMENTATION

Based on the proposed method, a prototype knowledge map system is being implemented in a knowledge management project for a Chinese manufacturing enterprise. The enterprise will be referred to as 'HFC' due to confidentiality. HFC is a large state-owned aviation industrial limited company in northeast China. Its main business scope covers the design and development of aeroplans. Design is a knowledge-intensive activity. After more than 50 years of development, HFC has accumulated a large number of unordered knowledge materials, which leads to the ineffectiveness of knowledge searching and using, so it is an urgent task for HFC to organize its knowledge resources and construct its own knowledge map.

The system employs service-oriented architecture (Zhang et al., 2008). Individual tagging, work centre knowledge recommendation and organizational knowledge map construction are designed to be the three fundamental functions of the system.

Individual tagging is an important function in the system. It allows users to organize knowledge of interest during his work. When a user tags a document in the repository, the tagging action will be stored as a record and the tag will be saved as an attribute, as is shown in Figure 6. Figure 7 shows the user interfaces of individual tagging.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

We design a virtual collaboration environment, work centre, to manage the knowledge in a domain. Work centre is a platform where members can share opinions about knowledge classification application. On one hand, every member's tagging data will be collected and analysed, and a recommendation based on collaborative tagging will be given. On the other hand, an expert in this domain draws the domain knowledge map according to his/her expertise and experiences. The enterprise knowledge map is constructed when the topics in different domain are related to each other by similarity.

CONCLUSION

In this paper, we design a method for enterprise knowledge map construction based on social classification. Social classification is currently popular on the internet. We demonstrate the collaborative tagging will provide some pragmatic benefits for enterprise knowledge organization. Unlike other methods for knowledge map construction, the proposed method develops the enterprise knowledge map as a multilevel system including knowledge tagging on individual level, topic selection on the domain level and topic association on the inter-domain level. By virtue of the collaborative nature of social classification and the similarity of knowledge structure in a business domain, our method proposes that the knowledge structure in a business domain can be analysed and integrated from individual knowledge tagging. The enterprise knowledge map is then organized based on domain to display topics, knowledge resource and their relationships.

The project is still going on and further research is more challenging. With the implementation of the system, we can obtain more tagging data from users. It is necessary in further research to refine the algorithms in the above method and validate them with the real data.

ACKNOWLEDGEMENTS

This paper is supported by the National Natural Science Foundation of China under Grant No.70671007.

REFERENCES

Argote L, Beckman SL, Epple D. 1990. The persistence and transfer of learning in industrial settings. Management Science 36(2): 140-154.

Chase RL. 1997. The knowledge-based organization: an international survey. The Journal of Knowledge Management 1(1): 38-49.

Chen Y, Li L. 2006. Deriving information from CRM for knowledge management-a note on a commercial bank. Systems Research and Behavioral Science 23(2): 141-146.

Chen H, Schuffels C, Owig R. 1996. Internet categorization and search: a machine learning approach. Journal of Visual Communications and Image Representation Science 7(1): 88-102.

Darr ED, Argote L, Epple D. 1995. The acquisition, transfer and depreciation of knowledge in service organizations: productivity in franchises. Management Science 41(11): 1750-1762.

Davenport T, Prusak L. 1998. Working Knowledge: How Organizations Manage What They Know. Harvard Business School Press: Boston.

Drabenstott KM, Weller MS. 1996. The exact-display approach for online catalog subject searching. Information Processing & Management 32(6): 719-745.

Eppler MJ. 2001. Making knowledge visible through intranet knowledge maps: concepts, elements, cases, In OT Proceedings of HICSS-2001, the 34th Hawaii International Conference on System Sciences.

Gruber T. 2005. Ontology of Folksonomy: A Mash-up of Apples and Oranges. The First on-Line conference on Metadata and Semantics Research. http://tomgruber.org/writing/ontology-of-folksonomy.htm [November 2005].

International Organization for Standardization. 2006. The Topic Maps Reference Model. http://www.isotopicmaps.org/tmrm/[2006].

Li L. 1999a. Proposing an architectural framework of hybrid knowledge-based system for production rescheduling. Expert Systems 16(4): 273-279.

Li L. 1999b. Knowledge-based problem solving: an approach to health assessment. Expert Systems with Applications 16(1): 33-42.

Li L, Zhao X. 2006. Enhancing competitive edge through knowledge management in implementing ERP systems. Systems Research and Behavioral Science 23(2): 129-140.

Maier R. 2004. Knowledge Management System: Information and Communication Technologies for Knowledge Management. Springer Verlag: Berlin.

Marwick AD. 2001. Knowledge management technology. IBM Systems Journal 40(4): 814-830.

Massicotte M. 1988. Improved browsable displays for online subject access. Information Technology and Libraries 7(4): 373-380.

Mathes A. 2004. Folksonomies-Cooperative Classification and Communication through Shared Metadata. http://www.adammathes.com/academic/computer_mediated-communication/ folksonomies.html_[December 2004].

Mertens K, Heisig P, Vorbeck J. 2003. Knowledge Management: Concepts and Best Practices. Springer: Berlin. Novak JD, Gowin DB. 1984. Learning How to Learn. Cambridge University Press: New York.

Ong T, Chen H, Sung W, Zhu B. 2005. Newsmap: a knowledge map for online news. Decision Support Systems 39(4): 583-597.

Pepper S. 2002. The TAO of Topic Maps: finding the way in the age of infoglut http://www.ontopia.net/topicmaps/materials/tao.html [April 2002].

Qi J, Xu L, Shu H, Li H. 2006. Knowledge management in OSS-an enterprise information system for the telecommunications industry. Systems Research and Behavioral Science 23(2): 177-190.

Sahraoui HA, Lounis H, Melo W, Mili H. 1999. A concept formation based approach to object identification in procedural code. Automated Software Engineering Journal 6(4): 387-410.

Studer R, Benjamins VR, Fensel D. 1998. Knowledge engineering: principles and methods. Data & Knowledge Engineering 25(1-2): 161-197.

Vail EF. 1999. Knowledge mapping: getting started with knowledge management. Information Systems Management 16(4): 16-23.

Warfield JN. 2007. Systems science serves enterprise integration: a tutorial. Enterprise Information Systems (1) 2: 235-254.

Warfield J. 2008. Seven challenges for information system designers. In Frontiers in Enterprise Integration, Xu L (ed.). Taylor & Francis: Boca Raton, FL, USA; 1-3. Wiig KM. 1997. Knowledge management: an introduction and perspective. The Journal of Knowledge Management 1(1): 6-14.

Xu L. 2000. The contribution of systems science to information systems research. Systems Research and Behavioral Science 17: 105-116.

Xu L, Wang C, Luo X, Shi Z. 2006. Integrating knowledge management and ERP in enterprise information systems. Systems Research and Behavioral Science 23(2): 147-156.

Yang B, Li L. 2001. Development of a KBS for managing bank loan risk. Knowledge-based Systems 14: 299-302.

Yang Y, Pedersen JO. 1997. A comparative study on feature selection in text categorization. In Proceedings of ICML-97, 14th International Conference on Machine Learning 412-420.

Yogesh M. 2000. Knowledge management and new organization forms: a framework for business model innovation. Information Resources Management Journal 13(1): 5-14.

Zhang T, Ying S, Cao S, Zhang J. 2008. A modeling approach to service-oriented architecture. Enterprise Information Systems 2(3): 239-257.

Lu Liu (1) *, Jing Li (2) and Chenggong Lv (1)

(1) School of Economics and Management, BeiHang University, Beijing 100083, China

(2) School of Economics and Business Administration, Beijing Normal University, Beijing 100875, China

* Correspondence to: Lu Liu, School of Economics and Management, BeiHang University, Beijing 100083, China.

E-mail: liulu@buaa.edu.cn
Table 1. A knowledge-tag matrix of an individual employee

           [t.sub.1]   [t.sub.2]   [t.sub.3]   [t.sub.4]

[d.sub.1]      1           0          1            0
[d.sub.2]      0           1          0            0
[d.sub.3]      0           1          1            1
[d.sub.4]      0           0          1            0
[d.sub.5]      1           1          1            1
[d.sub.6]      1           1          0            1

Table 2. All tagging records of a domain

                [d.sub.1]              [d.sub.2]

[p.sub.1]   [t.sub.1], [t.sub.3]         [t.sub.2]
[p.sub.2]      [t.sub.1]           [t.sub.1], [t.sub.3]
[p.sub.3]   [t.sub.1], [t.sub.3]   [t.sub.1], [t.sub.2], [t.sub.5]
[p.sub.4]   [t.sub.2], [t.sub.3]   [t.sub.2], [t.sub.3]
[p.sub.5]   [t.sub.1], [t.sub.3]

                    [d.sub.3]                  [d.sub.4]

[p.sub.1]   [t.sub.2], [t.sub.3], [t.sub.4]    [t.sub.3]
[p.sub.2]   [t.sub.2], [t.sub.3]
[p.sub.3]                                      [t.sub.3]
[p.sub.4]          [t.sub.2]                   [t.sub.3]
[p.sub.5]

                     [d.sub.5]

[p.sub.1]   [t.sub.1], [t.sub.2], [t.sub.3], [t.sub.4]
[p.sub.2]
[p.sub.3]   [t.sub.1], [t.sub.2], [t.sub.5]
[p.sub.4]            [t.sub.2]
[p.sub.5]   [t.sub.1], [t.sub.2]

                  [d.sub.6]                    [d.sub.7]

[p.sub.1]   [t.sub.1], [t.sub.2], [t.sub.4]
[p.sub.2]         [t.sub.4]                    [t.sub.1], [t.sub.4]
[p.sub.3]
[p.sub.4]                                      [t.sub.4]
[p.sub.5]

            [d.sub.8]                    [d.sub.8]

[p.sub.1]
[p.sub.2]                            [t.sub.4], [t.sub.5]
[p.sub.3]   [t.sub.2], [t.sub.5]        [t.sub.5]
[p.sub.4]                               [t.sub.5]
[p.sub.5]

Table 3. Knowledge-tag matrix of domain A

           [t.sub.1]   [t.sub.2]   [t.sub.3]   [t.sub.4]  [t.sub.5]

[d.sub.1]      4           1           4           0          0
[d.sub.2]      2           3           2           0          1
[d.sub.3]      0           3           2           1          0
[d.sub.4]      0           0           3           0          0
[d.sub.5]      3           4           1           1          1
[d.sub.6]      1           1           0           2          0
[d.sub.7]      1           0           0           2          0
[d.sub.8]      0           1           0           0          1
[d.sub.9]      0           0           0           1          3
COPYRIGHT 2009 John Wiley & Sons, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research Paper
Author:Liu, Lu; Li, Jing; Lv, Chenggong
Publication:Systems Research and Behavioral Science
Article Type:Report
Geographic Code:1USA
Date:Mar 1, 2009
Words:4987
Previous Article:Analysing relational benefits in e-business environment from behavioural perspective.
Next Article:A system dynamics approach to competitive strategy in mobile telecommunication industry.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |