An Ontology-Driven Knowledge Management System

Used in the Patent Library

Wei Ding

, Yongji Liu

and Jianfeng Zhang

China Defense Science and Technology Information Center, 100036, Beijing, China

National University of Defense Technology, 410073, Changsha, China

Keywords: Ontology-Driven, Knowledge Management, Patent Library.

Abstract: The introduction of the ontology-driven mechanism brings new opportunities for the knowledge

management. In this paper, we describe the results of our continuing work, by researching on the structure

design of the ontology-driven knowledge management system, and propose three stages in the system model,

i.e., the knowledge acquisition, the knowledge organization and knowledge application. Based on the

proposed model, we choose the technology novelty consulting system in our patent library as the platform to

perform the patent knowledge management and the experiment results validate the feasibility and validation

of our system model.

1 INTRODUCTION

The principle of "clear and precise description

specification of the domain conceptual model for

revealing the nature of the object" is referred to the

ontology (Razmerita, 2003). The introduction of the

ontology-driven mechanism and semantic web has a

significant impact on the knowledge management,

and the ontology based models are becoming a

straightforward choice of the future knowledge

management system. Online patent information and

document inspection services would become a

suitable application for such system.

The ontology-driven mechanism provides the

theoretical principle and technical support for the

knowledge organization, and the semantic web

draws a feasible blueprint for knowledge web

applications (Davies, 2007). The relevant researches

are growing fast recently, e.g., (Zhou, 2009; Yu,

2009; Bian, 2004) for building ontology-driven

models, (Xia, 2014; Dieng-Kuntz, 2006; Chau, 2007;

Cheng, 2009; Dong, 2006) for developing

application systems, and (Sureephong, 2007) for

introducing a case study used in the industry domain.

Based on these researches, three categories can be

drawn for describing the ontology-driven knowledge

management model: ① It is based on the knowledge

flow to build the components of the system model.

Such method focuses on the knowledge acquisition,

knowledge storage and knowledge re-utilization

(Huang, 2005), where the ontology is the core

technique to link all the nodes along the knowledge

flow (Chen, 2003). Examples of this kind include

(Jiang, 2009; Zhan, 2010). ② It is based on the

knowledge composition to describe the layers of the

system structure. Such ontology-driven knowledge

management system contains an application layer, a

core integration layer, a resource layer, etc. One

may utilize the domain ontology on the integration

layer, the notated documents on the application layer,

and the processed knowledge in the document

system or the database on the resource layer

(Benjamins, 1998). An extended example of this

kind can be find in (Chau, 2008), where the

integration layer is divided into the description layer

and the object layer, such that the ontology is

recognized to be either the concept library or the

object library on the description layer, while users

utilize the specific knowledge on the object layer. ③

It is from the angle of the system development that

the functional components of the system model are

described. Users utilize the knowledge resource

through the intermediate procedures such as

semantic analysis and ontology inference (Sun,

2009). A realistic example is witnessed in (Maedche,

2003), where the components, including user

interfaces, knowledge management, knowledge

model, workflow management and intelligent

application system, are developed while the

ontology is stored in the knowledge model.

248

Ding, W., Liu, Y. and Zhang, J..

An Ontology-Driven Knowledge Management System Used in the Patent Library.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 3: KMIS, pages 248-253

ISBN: 978-989-758-158-8

The common drawback of the aforementioned

literature is that their main focus is put on the

function of the ontology in the system; however no

clear and complete model for the knowledge

management system is concluded. Besides, there are

barely few studies for utilize the ontology-driven

knowledge management on the patent information.

In this paper, we propose a ontology-driven and

semantic-web-oriented structure to build the

knowledge management system model, and then

describe the detailed components and its workflow.

Such model is applied to the knowledge

management in our patent database and a case study

indicates the feasibility and validation of our system

model.

2 ONTOLOGY-DRIVEN

KNOWLEDGE MANAGEMENT

SYSTEM MODEL

In this section, we give a systematic view of the

proposed knowledge management system as

depicted in Fig.1. The whole model is divided into

three parts, i.e., the knowledge acquisition, the

knowledge organization and knowledge application,

similar as in (Antezana, 2009). At the stage of the

knowledge acquisition, the information from

different sources is first pre-processed, then the

specific objective knowledge (including the instant

objects, object relationship and logic rules) is

attained through the machine learning or manual

extraction, and afterwards they are generalized to

form the abstract knowledge (including concepts,

concept relationship, axioms and inference rules). At

the stage of the knowledge organization, the

acquired knowledge is described in an object-

oriented form (e.g., XML) to construct the domain

knowledge ontology library which is used to guide

the semantic annotations for attaining new object

libraries as the extension of the original knowledge

ontology. In such a way, a knowledge map can be

built based on the effective organization of the

domain knowledge. At the stage of the knowledge

application, services such as the knowledge topology

depiction, the knowledge search and the knowledge

innovation are available for users, wherein the

knowledge innovation also gives a feedback to

extend the knowledge ontology. Meanwhile, the

semantic web shall be utilized to build a network

platform to offer all the services.

2.1 Knowledge Acquisition

The so-called domain-knowledge mainly comes

from four sources. ① The domain information.

Literature is the main form to store the domain

information, which is increasingly expressed as texts

in the network service of a patent information library.

② Experts. Domain knowledge also exists in the

memories of the experts in this domain, some of

which may not be recorded as texts yet. ③ Existing

knowledge. Many researchers have contributed to

build domain knowledge libraries or the domain

ontology, e.g., in (Dong, 2006; Yan, 2007). ④

External information. It means the other sources

relevant to the studied domain. One example is like

that, for the knowledge “Method and means for

creating anti-gravity illusion are patented”, its

relationship with the song “Smooth Criminal” is not

included in this knowledge. However, it is good to

know that this song helps people to wildly remember

such anti-gravity illusion performed by the patent

inventor Michael Jackson. Information from

Figure 1: Ontology-driven knowledge management system structure.

An Ontology-Driven Knowledge Management System Used in the Patent Library

249

different sources is needed to be processed first into

a unified form and then transformed into the domain

knowledge. More specifically, the domain

knowledge is categorized into two forms. ①Abstract

knowledge. It indicates those summarized

descriptions based on objective facts by abstracting

the common characteristics, which is mainly

reflected as concepts, principles, rules, etc. For

example, “a person applies a patent” summarizes a

certain kind of behaviors, and it is also usefully to

guide semantic annotations of objective facts of all

patent applications. ② Objective knowledge. It

indicates those detailed descriptions of objective

facts. For example, “Method and means for creating

anti-gravity illusion is patented by Michael Jackson”

is an objective knowledge. The “Michael Jackson” is

the objective description of the inventor, and the

patent itself is detailed by “Method and means for

creating anti-gravity illusion”, while “patented”

shows their relationship. The objective knowledge is

the sole resource to attain the abstract knowledge,

which can be obtained by machine learning or

manual extraction aided by experts.

2.2 Knowledge Organization

It is significant to organize the acquired knowledge

into a domain knowledge library for further

utilization or knowledge accumulation. Typical

methods such as the thesaurus (Dong, 2006; Yan,

2007) are widely adopted for knowledge

organization. However, the complicated semantic

relationship between knowledge is beyond the

capabilities of these methods. Along the

development of the ontology-driven mechanism in

the knowledge engineering, the object-oriented

manner is more suitable for organizing knowledge,

where the object properties are used to describe the

knowledge relationship. Fig.2 depicts an example of

such detailed description in the form of XML,

wherein the “Person” and “Patent” is the concept

while “Apply” and “Applied” is their properties,

respectively.

The synthesized domain knowledge can be

further used to guide for semantic annotations (Uren,

2006; Zhang, 2002), which mainly have two

methods.①manual annotating. Experts may label the

domain knowledge based on the ontology definition,

which is usually time-consuming. ② automatic

annotating. It is typical to adopt a language

processing technology to automatically perform the

semantic analysis and complete the annotation. No

matter manual annotating or automatic annotating,

many computer tools can be found (Zou, 2004) to

accelerate the efficiency and accuracy.

Figure 2: Example of knowledge organization description.

2.3 Knowledge Application

Based on the researches on existing literature, we

can summarize that there are three areas in the

knowledge applications, i.e., the knowledge map, the

knowledge searching and the knowledge innovation.

The semantic web, as the future network platform, is

considered as an effective tool for the knowledge

management and its service.

Figure 3: A segment illustration of the knowledge map.

1) Knowledge Map

The knowledge map is originally extended from the

cognitive map introduced in (Brookes, 1981), and

afterwards it becomes a visible knowledge

management method which is widely adopted in

many domains (Eppler, 2008; Rao, 2012; Semenza,

2012). Knowledge map is useful to perform the

knowledge navigation and progressive guide service.

As shown in Fig.3, from “Person 1”, one may extend

more relevant information, e.g., all his applied

patents in the intelligence analysis and the

knowledge about his co-applicants, through the

visible browsing guide provided by the knowledge

map. The knowledge ontology would include all the

knowledge in this domain, however the concealed

knowledge, e.g., the relevant information, is hard to

be attained and thus needed to be revealed by

knowledge inference and data mining. In Fig. 4, the

correlation coefficient between “Person A” and

KMIS 2015 - 7th International Conference on Knowledge Management and Information Sharing

250

“Person B” is illustrated by the knowledge map as

inspired by (Yan, 2007). Normally, the correlation

information is not included in the domain ontology

but can be attained by the knowledge inference.

Figure 4: An illustration of the knowledge inference using

the knowledge map.

2) Knowledge Searching

About the knowledge searching, there is no

determined definition. In a broad sense, it imitates

the intelligent cognitive methods of the human being

(Zhang, 2003). In this paper, we follow (Cheng,

2011) to consider it as an information retrieval

technique. More specifically, we adopt the semantic

searching technique introduced in (Kiryakov, 2004)

to improve the searching quality, allowing users to

attain required resources more precisely and more

conveniently. Its function can be described as the

follows. ① Indexing service used in the domain

knowledge resource. For example, one user would

not only attain the required patent information, but

also other relevant knowledge according to the

semantic matching proposed in (Jiang, 2009). ②

Extended searching in other domain resources based

on the knowledge information. After building the

domain ontology, it can also be used for searching

the same or relevant resources. For example, the

knowledge searching in our technology novelty

consulting system would also refer to other domain

ontology, e.g., in IEEE database.

3) Knowledge Innovation

The knowledge innovation is an important method to

enhance the maturity and extension of the domain

knowledge ontology. For the example as shown in

Fig. 4, the correlation between two persons can be

attained by knowledge inference. Suppose there is a

fixed threshold, the knowledge about one person

shall be involved into the domain knowledge

ontology of the other person in Fig. 4, when their

correlation coefficient is larger than the threshold. In

this manner, the original knowledge ontology is

extended. The knowledge innovation mode is

summarized as follows. ①the knowledge innovation

based on the internal knowledge inference as shown

in Fig. 5(a). It reveals the concealed knowledge

through methods such as comprehensive and

comparative analysis, logic reasoning, machine

learning and data mining, etc. For example, from the

frequency of subject keywords, the research trends

of the hot topics can be concluded. ②the knowledge

innovation by considering the external knowledge as

shown in Fig. 5(b). Such example is mentioned

before, which is that Michael Jackson’s patent

“method and means for creating anti-gravity illusion

are patented” is widely known related with his song

“Smooth Criminal”.

3 CASE EXPERIMENT

We adopt the aforementioned system model in our

patent knowledge management and embed it in the

technology novelty consulting system in our patent

library. Fig.6 describes the knowledge management

workflow in this case experiment, for which we

choose the defense patent documents applied by a

Figure 5: Ontology-Driven Knowledge Innovation.

An Ontology-Driven Knowledge Management System Used in the Patent Library

251

Figure 6: Patent Ontology-Driven Knowledge Management Workflow.

chosen institution from 2006 to 2014 as the data

resource. These data exist in a text database form

stored in our patent library. Based on the features of

these patents, we perform the knowledge acquisition.

Figure 7: Correlation Cloud Map of Selected 11 Patents.

For organizing the knowledge, we first abstract

the acquired knowledge to build the ontology

concept model, which is described in a manner

illustrated in Fig. 2. Then the semantic annotation is

completed by performing object extraction and

setting the object properties. For these knowledge,

the correlation cloud map of the selected 11 patent

objects is generated using the semantic analysis

described in (Yan, 2007) and the map is shown in

Fig. 7. In comparison, we also apply (Yan, 2007) to

generate the cloud map in our case. The numerical

results show that our method is 1.21 times faster.

One reason is that the adopted model in this paper

has already given clear and complete workflow for

the knowledge management system, which acts like

a certain pre-progress procedure to help accelerate

the knowledge management efficiency. Moreover,

the feature of patents can not be used in the of the

knowledge management system of (Yan, 2007) but

we can in this paper.

Figure 8: Fields with most patents in the test database.

Furthermore, we analyze the keyword frequency

in the tested patent database and depict Fig. 8,

illustrating the fields with most patents in the test

database. As a knowledge application, this figure

indicates that the major expertise of the chosen

institution lies in the field of the aviation and

aerospace.

4 CONCLUSIONS

In this paper, a complete structure of ontology-

driven knowledge management system model is

introduced, and the detailed workflow is described.

Based on the proposed model, a case study applied

to the patent knowledge management is performed

to validate the system model.

However, this work is still in progress, since it

still lacks efforts on the semantic analysis on the

patent information, the ontology-driven searching

and the patent-evaluating service. These functions

have been already known highly required by the

users, and thus will be our tasks in future.

KMIS 2015 - 7th International Conference on Knowledge Management and Information Sharing

252

ACKNOWLEDGEMENTS

We thank Dr. Tao Xu for his invaluable advice and

suggestions, who is in part supported by China

Postdoctoral Science Foundation (2015M570230)

and Tianjin Enterprise-Postdoctoral Fund for

Selected Innovation Program, at University-

Enterprise joint postdoctoral station between

Tsinghua University and Tianjin Zhonghuan

Electronic & Information (Group) Co., Ltd.

REFERENCES

Razmerita, L., et al, 2003. Ontology-based user modeling

for knowledge management systems, Proceedings of

the User Modeling.

Davies, J., et al, 2007. Semantic-Web-based knowledge

management, Internet Computing.

Zhou, Y., et al, 2009. The design and implementation of

digital archives knowledge management system in e-

government domain based on ontology, Library and

Information Service.

Yu, S., et al, 2009. Research on multiuser-oriented

medical knowledge management model based on

ontology, Journal of Medical Intelligence.

Bian, W., et al, 2004. Ontology-based knowledge

management modeling of scientific research,

Application Research of Computers.

Xia, L., et al, 2014. Research on e-government knowledge

management based on ontology, Information Science.

Dieng-Kuntz, R., et al, 2006. Building and using a medical

ontology for knowledge management and cooperative

work in a health care network, Computers in Biology

and Medicine.

Chau, K., et al, 2007. An ontology-based knowledge

management system for flow and water quality

modeling, Advances in Engineering Software.

Cheng, H., et al, 2009. An ontology-based business

intelligence application in a financial knowledge

management system, Expert Systems with

Applications.

Dong, H., et al, 2006. Research on the ontology-based

retrieval model of digital library, Journal of the China

Society for Scientific and Technical Information.

Sureephong, P., et al, 2007. An ontology-based knowledge

management system for industry clusters, IEEM'2007.

Huang, X., et al, 2005. Research of knowledge

management system based ontology, Science

Technology and Engineering.

Chen, R., et al, 2007. Research of knowledge mangement

based on ontology, Science Technology Association

Forum.

Jiang, Y., et al, 2009. Hotspots analysis of research on

ontology-based knowledge management, Computer

Knowledge and Technology.

Zhan, Y., et al, 2010. Knowledge management models

based on ontology, Wireless Internet Technology.

Benjamins, V. R., et al, 1998. Knowledge management

through ontologies, Proceedings of PAKM98.

Sun, W., et al, 2009. Anontology-based manufacturing

design system, Information Technology Journal.

Maedche, A., et al, 2003. Ontologies for enterprise

knowledge management, Intelligent Systems.

Antezana, E., et al, 2009. Biological knowledge

management: The emerging role of the semantic Web

technologies, Briefings in Bioinformatics.

Dong, H., et al, 2006. Research on the ontology-based

retrieval model of digital library(Ⅲ)-History domain

ontology building, Journal of the China Society for

Scientific and Technical information.

Yan, D., et al, 2007. Visualized searches of domain

ontologies, Journal of Library Science in China.

Uren, V., et al, 2006. Semantic anannotation for

knowledge management: Requirements and a survey

of the state of the art, Science, Services and Agents on

the World Wide Web.

Zhang, X., 2002. Semantic web and semantic-based

networked information retrieval, Journal of the China

Society for Scientific and Technical Information.

Zou, L., et al, 2004. Comparison and analysis of semantic

anannotation tools based on ontology, Journal of

Computer Applications.

Brookes, B. C., et al, 1981. The foundations of

information science Part IV, Journal of Information

Science.

Eppler, J., et al, 2008. A process-based classification of

knowledge maps and application examples,

Knowledge and Process Management.

Rao, L., et al, 2012. Building ontology based knowledge

maps to assist business process re-engineering,

Decision Support Systems.

Semenza, J., et al, 2012. Knowledge mapping for climate

change and food-and waterborne diseases, Critical

Reviews in Environmental Science and Technology.

Zhang, Y., et al, 2003. On knowledge retrieval and

information retrieval, Journal of Library Science in

China.

Cheng, H., et al, 2011. Research on development of

studying on domestic knowledge retrieval, Library

and Information Service.

Kiryakov, A., et al, 2004. Semantic anannotation, indexing,

and retrieval, Science, Services and Agents on the

World Wide Web.

Jiang, X., et al, 2009. Learning and inferencing in user

ontology for personalized semantic Web search,

Information Sciences.

An Ontology-Driven Knowledge Management System Used in the Patent Library

253