Knowledge Fusion

in Context-Aware Decision Support Systems

Alexander Smirnov

1,2

, Tatiana Levashova

and Nikolay Shilov

1,2

St. Petersburg Institute for Informatics and Automation of the Russian Academy of Sciences,

39, 14th line, St. Petersburg, Russia

ITMO University, 49, Kronverkskiy av., St. Petersburg, Russia

Keywords: Context-Aware Decision Support, Knowledge Fusion, Emergency Response, Response Actions Planning.

Abstract: The paper discusses knowledge fusion processes with reference to context-aware decision support. It

extends the previous research work on context-based knowledge fusion patterns. The contribution of this

paper is service-oriented implementation of the context-aware decision support system for emergency

management. This system was used in the previous work as the basis for revealing the knowledge fusion

patterns. The presentation is accompanied by examples from a fire response scenario.

1 INTRODUCTION

The decision support systems heavily rely upon

large volumes of data, information, and knowledge

arriving from different sources. Whereas several

years ago data fusion used to be the main technology

integrating data and information from multiple

sources within any decision support system, today

the focus of data fusion has changed to knowledge

fusion. The objective of knowledge fusion is to

integrate information and knowledge from multiple

sources into some new common knowledge that may

be used for decision making and problem solving or

may provide a better insight and understanding of

the situation under consideration (Holsapple &

Whinston 1986; Preece et al. 2001).

In the present research, an emergence of new

knowledge is considered as the distinguishing

feature of the knowledge fusion processes. Any

sources of data, information, and knowledge

involved in the fusion processes are referred to as

knowledge sources.

In the research, the knowledge fusion problem is

considered applying to decision support systems

intended for usage in dynamic environments. Such

systems have to be context-aware in order to control

environmental changes, to adapt to the current

situation, and to avoid information overload.

Context enables to decrease the volumes of available

information and knowledge to the information and

knowledge relevant to or "useful" in the current

situation.

The research extends the preceding research

work on context-based knowledge fusion patterns

(Smirnov et al. 2013) revealed in a context-aware

decision support system (CADSS) for emergency

management. The contribution of this paper is the

system implementation. The motivation is to

illustrate knowledge fusion processes by a practical

example avoiding descriptions of technical details.

The rest of the paper is structured as follows. The

following Section presents the main research areas

dealing with knowledge fusion and introduces dif-

ferent results that knowledge fusion can produce.

Section 3 outlines the conceptual framework of the

CADSS, discusses the place of knowledge fusion in

the system's scenario, and presents some implemen-

tation issues by example of a fire response scenario.

The Conclusion emphasizes the role of knowledge

fusion for up-to-date decision support systems.

2 KNOWLEDGE FUSION

Recently, many research focus on knowledge fusion.

Depending on the application domains the focus of

knowledge fusion becomes more specific.

Knowledge fusion technology supports intelligent

query-answering systems aiming at providing the

users with nonrecurring, consistent, unambiguous

186

Smirnov A., Levashova T. and Shilov N..

Knowledge Fusion in Context-Aware Decision Support Systems.

DOI: 10.5220/0005034801860194

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2014), pages 186-194

ISBN: 978-989-758-049-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

answers to their query under redundancy/lack of

information (e.g., Dumais et al. 2002; Nengfu et al.

2012; Preece et al. 1999). Some researchers deal

with knowledge fusion of multiple knowledge

sources to build a new knowledge base (Craven et

al. 2000; Gou et al. 2005; Kuo et al. 2003). Several

knowledge fusion efforts are devoted to knowledge

integration with the problem solving purposes

(Carvalho et al. 2013; Preece et al. 1999; Smirnov et

al. 2005b). A number of knowledge fusion

applications deal with fusing information to obtain a

model of the situation, assess it, and predict its

development (Blasch et al. 2013; Boury-Brisset

2001; Erlandsson et al. 2010; Golestan et al. 2013;

Pengpeng Liang et al. 2013; Sanchez et al. 2014).

Context plays an important role in all the above

mentioned approaches.

As it is said in the Introduction, knowledge

fusion is characterized by the creation of new

knowledge. The analysis of a number of knowledge

fusion studies (above mentioned and additionally to

them (Besnard et al. 2012; Grebla et al. 2010;

Jonquet et al. 2011; Lee 2007; Lin & Lo 2010;

Preece et al. 2001; Roemer et al. 2001; Smirnov et

al. 2005b)) has revealed the following kinds of new

knowledge obtained as the knowledge fusion results:

 new knowledge created from data/information;

 a new type of knowledge;

 a new problem solving method or a new idea how

to solve the problem;

 a new knowledge about the conceptual scheme;

 new capabilities/competencies of a knowledge

object (an object that produces or contains

knowledge);

 a solution for the problem;

 a new knowledge source.

The CADSS for the emergency management

domain was investigated with the object to reveal in

it the listed above knowledge fusion results and to

identify the processes eventuating in them.

3 KNOWLEDGE FUSION IN

CADSS

The CADSS for emergency management is intended

to support decisions on planning emergency

response actions. The system scenario follows two

main phases: preliminary and executive (Figure 1).

These phases comprise several stages; at some of

them knowledge fusion processes occur.

3.1 Preliminary Phase

At the preliminary phase, an application ontology,

which describes knowledge of the emergency

management domain, is built. This ontology

represents non-instantiated knowledge of two types:

domain and problem solving. The domain

constituent of the application ontology represents

various types of emergency events and knowledge

that can be used to describe situations caused by

such events; the problem-solving constituent

represents problem-solving knowledge that may be

required to solve the problem of planning the

emergency response actions. The application

ontology is the result of the integration of multiple

Environmental

knowledge sources

Application

ontology

Abstract context

Alternative

decisions

Preliminary phase Executive phase

Contextual

knowledge sources

Knowledge source

network

Decision

making

Operational context

Figure 1: Context aware decision support.

KnowledgeFusioninContext-AwareDecisionSupportSystems

187

required to solve the problem of planning the

emergency response actions. The application

ontology is the result of the integration of multiple

pieces of different ontologies provided by various

ontology libraries and is semi-automatically created

by domain experts (Smirnov et al. 2005a). At the

stage of application ontology building, the

knowledge fusion result manifests as a new ontology

created from multiple source ontologies, at that this

ontology is of a new type. The new type comes out

of fusion of knowledge of the domain type and

knowledge of the problem solving type. Therefore,

the stage of application ontology building produces

two sorts of knowledge fusion results like a new

knowledge source created from multiple sources and

a new type of knowledge.

The application ontology is formalized through

constraints. It is represented by sets of classes, class

attributes, attribute domains, and constraints

(Smirnov et al. 2003). The ontology specified in this

way corresponds to a (non-instantiated) object-

oriented constraint network.

The application ontology of the emergency

management domain is not represented here because

of its largeness.

3.2 Executive Phase

The executive phase concerns context-aware support

of the decision maker with alternative decisions,

decision making, decision implementation, and

archiving. Context model is used to represent

knowledge about the emergency situation. Abstract

context and operational context represent the

situation at the first level and the second level,

respectively (Figure 1).

3.2.1 Abstract Context Creation

The application ontology serves as the basis for

abstract context creation. This context captures from

the application ontology the knowledge pieces

relevant to the current emergency situation and

significantly reduces the amount of knowledge

represented in the ontology. The abstract context is a

non-instantiated object-oriented constraint network

just like the application ontology. Both components

(domain and problem solving knowledge) making

up the application ontology are presented in the

abstract context. From the knowledge fusion

perspective the abstract context is a new knowledge

source created through integration of multiple pieces

of the single knowledge source (the application

ontology).

Figure 2 presents the abstract context created for

a fire situation. The created context, along with other

issues, specifies that in the fire situation the services

provided by hospitals, emergency teams and fire

brigades are required. The emergency teams, and

fire brigades are mobile resources; they can use

ambulances, fire engines, and special-purpose

helicopters for transportation.

In the figure, the problem-solving knowledge

specified in the abstract context is collapsed in the

“Emergency response” class. Partly, this class is

shown expanded on the right. The class specifies the

following problems:

 select feasible hospitals, emergency teams, and

fire brigades;

 determine feasible transportation routes for

ambulances and fire engines depending on the

transportation network and traffic situation;

 calculate the shortest routes for transportation of

the emergency teams by ambulances and fire

brigades by fire engines;

 produce a set of feasible response plans for

emergency teams, fire brigades, and hospitals.

The ontology inference supports the procedure of

knowledge integration into the abstract context. This

inference is based on the regularities of the ontology

Domain knowledge

Problem solving knowledge

Figure 2: Abstract context (a fragment).

Abstract context: new knowledge source created

from multiple knowledge pieces

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

188

representation formalism used. Consequently, new

ontology representation items (constraints, classes,

attributes) may appear in the abstract context. This is

the case of producing a new knowledge about the

conceptual scheme as the result of knowledge

fusion.

In the course of the abstract context for the fire

situation creation a new relationship between the

knowledge unrelated in the application ontology has

been inferred (Figure 3). The application ontology

specifies that a value for the argument representing

the current location of a transportation device serves

as an input argument of the routing method (1). The

class “mobile” representing a mobile resource and

the class “transportation device” are linked by a

functional relationship (2) stating that the location of

a mobile resource is the same as the location of the

transportation device this resource goes by. In the

abstract context a new functional relationship (3) has

been inferred. This relationship means that a value

for the attribute representing the current location of a

mobile resource serves as an input argument of the

routing method. In other words, values for the both

attributes representing the current location of a

transportation device or the current location of a

mobile resource can be used as one of the input

arguments by the routing method.

3.2.2 Operational Context Producing

The operational context is an instantiation of the

abstract context with the actual information. A

subset of all available environmental knowledge

sources is organized to instantiate the abstract

context. This subset is referred to as contextual

knowledge sources. The subset of the contextual

knowledge sources comprises knowledge sources

that can provide data values to create instances of

the classes represented in the abstract context or

solve the problems specified in it. The contextual

knowledge sources with the specified sequence of

their execution organize a knowledge source

network. Nodes of this network are knowledge

sources providing data values and/or solving the

problems; network arcs signify the order of the

nodes execution (Figure 4). In this figure,

knowledge sources indicated by the same numbers

are executed simultaneously.

The operational context reflects any changes in

information incoming from the knowledge source

network. The CADSS produces a special view for

the operational context so that the users (decision

makers) would be able to read and understand it.

The operational context is the result of the

intelligent fusion of heterogeneous data /

information from the contextual knowledge sources.

This context is a new knowledge created from the

environmental information, which is intended to be

used as the foundation for problem solving and

decision making. Moreover, the operational context

represents a new knowledge type, namely the

dynamic type.

Figure 5 presents a view for the operational

context (the big dot denotes the fire location).

3.2.3 Problem Solving

Acting

Mobile

Transportation

device

Mob_Location

TDev_Location

Route

computation

Mob_Location

Resource

Emergency

response

is-a

associat

ive

function

part-o

inferred relationship (3):

Route computation = F4(Mob_Location)

new knowledge about conceptual scheme

Figure 3: Inferred relationship.

(2): F1(Mob_Location) = F2(TDev_Location)

(1): Route computation = F3(TDev_Location)

(1)

(2)

(3)

Fire location

Locations and

availabilities

of hospitals

Route

availabilities

Selection of

possible ETs

and FBs

Shortest route

calculation

Producing

a set of fire

response plans

Quantity of

ETs and FBs

Number

of victims

Weather

conditions

Road

locations

Availabilities

of ETs and

FBs

Locations of

ETs and FBs

Figure 4: Knowledge source network.

ET – Emergency Team;

FB – Fire Brigade

KnowledgeFusioninContext-AwareDecisionSupportSystems

189

The knowledge source network solves the planning

problem embedded in the operational context as a

constraint satisfaction problem. Consequently a set

of alternative emergency response plans each

corresponding to a decision that can be made in the

current situation is received. An emergency response

plan is a set of emergency responders with required

helping services, schedules for the responders’

activities, and transportation routes for the mobile

responders. The process of problem solving is a

process of fusion of knowledge from various sources

in problem solving, which results in a solution. The

solution forms a new knowledge type that is non-

ontological knowledge.

Figure 5 presents a plan for actions for emer-

gency teams, fire brigades, and hospitals (the dotted

lines designate the routes proposed for the transpor-

tations of the emergency teams and fire brigades).

3.2.4 Decision Making and Decision

Implementation

The decision maker chooses one plan from the set of

alternative ones by selecting any plan from the set or

taking advantage of some predefined efficiency

criteria (e.g., minimal time of the response actions,

minimal cost of these actions, minimal time of

transportation to hospitals, etc.). The chosen plan is

considered to be the decision.

The decision is delivered to the emergency

responders included in the plan, i.e., to the actors

responsible for the plan implementation. They have

to approve the decision by confirming their

readiness to participate in the response actions (the

decision implementation). The emergency

responders are enabled to participate in the approval

procedure using any Internet-accessible devices. The

approval of the decision by the actors directly allows

ones to avoid hierarchical decision making, which is

time-consuming and is not good for emergencies.

If some of the emergency responders are not able

to participate in the plan then, in some cases, the

plan can be adjusted. The option of plan rejection is

provided for due to the rapidly changing emergency

situations – something may happen between the

moment when the decision is made and the time

when the emergency responders receive the plan.

The plan adjustment implies a redistribution of the

functions appertaining to the refused emergency

responders between the other plan participants.

In the CADSS, emergency responders are

represented by their profiles. If an emergency

responder agrees to participate in some activity that

its profile does not provide for then this profile is

extended with a new capability. This is the case of

fusion of implicit (unspecified capabilities)

knowledge and explicit (profile) knowledge. The

knowledge fusion result is gaining new capabilities /

competencies by a knowledge object.

3.2.5 Archival Knowledge Management

The decision, the abstract context, and the

operational context along with the knowledge source

network are saved in a context archive. The

operational context and the knowledge source

network are saved in their states at the instant of the

alternatives’ generation. The archived components

are the objects of archival knowledge management.

The archival knowledge management pursues

several goals (e.g., revealing user preferences,

grouping users with similar interests, decision min-

ing, etc.). With reference to knowledge fusion, the

main intention of the archival knowledge manage-

ment is inference of new knowledge based on the

accumulated one. For instance, new relations be-

tween the knowledge represented in the operational

contexts can be discovered based on a comparative

analysis of these contexts accumulated in the context

archive, i.e. based on fusion of the knowledge

represented by the operational contexts. Finding the

same instance in different operational contexts may

lead to revealing new relations for this instance.

Figure 5: Operational context fused with decision.

Operational context: new type of new knowledge

created from data/information

a problem solution –

new type of knowledge

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

190

The archival knowledge management enables to

reveal explicit knowledge from the hidden one and

new relations between originally unrelated

knowledge. The both outcomes are interpreted as a

new knowledge about the conceptual scheme, which

refine the existing representations.

For instance, the emergency team encircled in

Figure 5 participated in different emergency

response actions. Some operational contexts in

which this team appeared and then participated in

corresponding actions do not represent any instances

of the class Emergency response organization

specified in the abstract context. This suggests that

the emergency team is a part of one of the hospitals

represented in the operational contexts together with

this team. Based on the operational context (Figure

6) it can be judged that most probably the team is a

part of hospital 5 represented in this context since

the context does not represent any other hospitals

from Figure 5 except this one. Part-of relation

between the hospital 5 and the encircled emergency

team is the new revealed relation. This relation is an

outcome of inductive inference.

3.2.6 Abstract Context Reuse

Reuse of an abstract context in settings when the

available knowledge sources are not intended to

solve the problems specified in this context is a

reason to search for a new configuration of the

knowledge source network so that this new configu-

ration would be able to solve the specified problems.

Search for a new configuration implies search for

knowledge sources that can solve the problems using

the methods specified in the context as well as

search for alternative problem solving methods. A

basic condition for finding alternatives is an availa-

bility of knowledge sources that provide methods

that can be used to solve the specified problems.

If alternative methods have been found, they get

specified in the abstract context. That is, the abstract

context gets extended with the new knowledge about

the conceptual scheme. The knowledge source

network is reconfigured accordingly. The new

configuration gains new capacities.

Reuse of an abstract context can produce several

knowledge fusion outcomes. Namely, a new prob-

lem solving method can be found, a new knowledge

about the conceptual scheme can appear, and a new

configuration of the knowledge source network with

new capacities / competencies can be organized.

The abstract context (Figure 2) was reused in

settings where the knowledge source intended to

provide information about hospitals' locations

missed. The abstract context specifies the routing

problem as a hierarchy of methods one of which

('GetLocation') returns the current locations of

objects in the format of point coordinates on the

map. In the example under consideration it is

required to determine the locations of hospitals. The

method 'GetLocation' uses data from sensors.

The set of environmental knowledge sources

comprises no sensors dealing with static objects like

hospitals, and this set comprises some other sources.

One of them implements a method

('MedicalCareSuggestions') intended to make

recommendations what medical care organizations

can be used to access some specific medical service.

This source contains a database with information

about the hospitals. The method

'MedicalCareSuggestions' returns the hospitals'

addresses in an address format. The other source

implements the method ('Conversions') that converts

the address formats into the format of coordinates.

The successive execution of the methods

'MedicalCareSuggestions' and 'Conversions' is an

alternative way to calculate the hospital locations in

the format of coordinates.

In the abstract context the methods

'MedicalCareSuggestions' and 'Conversions' are not

specified as an alternative to the method

'GetLocation'. A set of constraints have to be

introduced to get this alternative explicitly specified.

This introducing leads to the extension of the

abstract context with new knowledge representation

items, that is new knowledge about the conceptual

scheme of the abstract context.

3.2.7 Service Oriented Architecture

The CADSS is implemented as a set of services. Its

architecture comprises three groups of services. The

first group is made up of core services responsible

for the registration of the services in the service

situation, i.e. for the creation of the abstract and

operational contexts.

Figure 6: History for emergency team.

part-of: new knowledge

about conceptual scheme

KnowledgeFusioninContext-AwareDecisionSupportSystems

191

Web-services forming the second group are

responsible for the generation of alternative plans for

actions and the selection of an efficient plan.

The third group comprises services responsible

for the representation of the environmental

knowledge sources and the emergency responders

and implementation of their functions.

The composition of services is applied for the

organization of the services representing the

contextual knowledge sources (contextual services)

to instantiate the abstract context. The abstract

context specifies an abstract workflow of the

required composite service. The services

communicate in terms of their inputs/outputs to

create a service execution sequence. If alternative

services available a set of sequences is created. A

specific alternative is chosen based on the principles

of maximum functionality, maximum access

interval, and minimum service weight.

Principle of maximum functionality. For the

services that implement several functions, usage of

one service implementing several functions is

considered more expedient than usage of several

services implementing the same functions

separately, i.e.

max



FNWF

for the n

service, where

WF – the set of functions the n

service implements,

– the set of functions

specified in the abstract context.

Principle of maximum access interval. If there

exist several services that are accessible over the

interval



Tt ,

at different time intervals





, then

selection of less number of services whose overall

access intervals cover the interval





Tt ,

considered to be more efficient. At that, the interval



Tt ,

must be fully covered by the intervals

t

(

t

– the i

access interval for the n

resource,





), i.e.





tTt





(

– number of

intervals

t

over the interval



Tt ,

for the n

resource,

– number of contextual services).

Principle of minimum weight. This principle is

used to evaluate the efficient selection of alternative

services. Alternative services are services

implementing the same functions but differing in

their locations, costs, etc. Weight of a service is

calculated as



CTN=W







1 , where

–

the service’s competence / reliability,





1,0N ;

– the average access time to the service relatively to

the access time that is maximum among the access

times for the alternative services;

– the cost of

information acquisition from the service relatively to

the cost of information acquisition that is maximum

among the costs for the alternative Web-services;







– relative importance of the parameters for

the particular service (

1









The above principles do not pretend to be

sufficient to produce optimal service networks; they

provide for efficient service network configuration.

4 CONCLUSIONS

Various knowledge fusion processes were

investigated. These processes produce different

knowledge fusion outcomes. The places of

knowledge fusion in the context-aware decision

support system for emergency management were

indentified. In this system, the knowledge fusion

outcomes were found. Knowledge fusion enables to

create a new knowledge. Such new knowledge

allows the systems to adapt to the current situations

efficiently. Therefore, it can be concluded that

efficiency of up-to-date decision support systems

depends heavily on their abilities to management of

knowledge fusion processes.

ACKNOWLEDGEMENTS

The present research was partly supported by the

projects funded through grants 12-07-00298,

13-07-12095, 14-07-00345, 14-07-00427 (the

Russian Foundation for Basic Research), the Project

213 (the Russian Academy of Sciences (RAS)), the

Project 2.2 (the Nano- & Information Technologies

Branch of RAS), and grant 074-U01 (the

Government of the Russian Federation).

REFERENCES

Besnard, P., Gregoire, E. & Ramon, S., 2012. Logic-based

fusion of legal knowledge. In: Proceedings of the 15th

international conference on information fusion,

Singapore, 9–12 July 2012, pp. 587–592.

Blasch, E., Herrero, J.G., Snidaro, L., Llinas J.,

Seetharaman, G., Palaniappan, K., 2013. Overview of

contextual tracking approaches in information fusion.

In: Proceedings SPIE8747, Geospatial InfoFusion III,

vol. 8747, Baltimore, Maryland, USA, 29 April 2013.

doi:10.1117/12.2016312.

Boury-Brisset, A.C., 2001. Towards a Knowledge Server

to support the Situation Analysis Process. In:

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

192

Proceedings of the 4th international conference on

information fusion, Montréal, Canada, 7–10 August

2001, viewed 24 June 2014, http://isif.org/fusion

/proceedings/fusion01CD/fusion/searchengine/pdf/Th

C23.pdf.

Carvalho, R.N., Laskey, K.B., Costa, P.C.G., Ladeira, M.,

Santos, L.L. & Matsumoto, S., 2013. Probabilistic

ontology and knowledge fusion for procurement fraud

detection in Brazil. In: F. Bobillo et al (eds.),

Uncertainty Reasoning for the Semantic Web II,

International Workshops URSW 2008–2010 held at

ISWC and UniDL. Lecture notes in computer science,

vol. 7123, pp. 19–40.

Craven, M., DiPasquo, D., Freitag, D., McCallum, A.,

Mitchell, T., Nigam, K. & Slattery, S., 2000. Learning

to construct knowledge bases from the World Wide

Web, Artificial Intelligence, vol. 118, pp. 69–113.

Dumais, S., Banko, M., Brill, E., Lin, J. & Ng, A., 2002.

Web question answering: is more always better? In:

Proceedings of the 25th annual international ACM

SIGIR conference on research and development in

information retrieval, Tampere, Finland, 11–15

August 2002, pp. 291–298.

Erlandsson, T., Helldin, T., Falkman, G. & Niklasson, L.,

2010. Information fusion supporting team situation

awareness for future fighting aircraft. In: Proceedings

of the 13th international conference on information

fusion, Edinburgh, UK, 26–29 July 2010, IEEE,

viewed 22 June 2014, http://ieeexplore.ieee.org

/stamp/stamp.jsp?tp=&arnumber=5712064.

Golestan, K., Karray, F., Kamel, M.S., 2013. High level

information fusion through a fuzzy extension to multi-

entity Bayesian networks in vehicular ad-hoc

networks. In: Proceedings of the 16th international

conference on information fusion, Istanbul, Turkey, 9–

12 July 2013, pp. 1180–1187.

Gou, J., Yang, J. & Chen, Q., 2005. Evolution and

evaluation in knowledge fusion system. In: J. Mira &

J.R. Alvarez (eds.), IWINAC 2005, International work-

conference on the interplay between natural and

artificial computation, Las Palmas de Gran Canaria,

Canary Islands, Spain, 15–18 June 2005. Lecture notes

in computer science, vol 3562, pp. 192–201.

Grebla, H.A., Cenan, C.O. & Stanca, L., 2010. Knowledge

fusion in academic networks, Brain: Broad Research

in Artificial Intelligence and Neuroscience, vol. 1, no.

2, viewed 17 June 2014, http://www.edusoft.ro-

/brain/index.php/brain/article/download/60/145.

Holsapple, C.W. & Whinston, A.B., 1986. Building blocks

for decision support systems. In: G. Ariav & J.

Clifford (eds.), New directions for database systems,

Ablex Publishing Corp., Norwood, pp. 66–86.

Jonquet, C., LePendu, P., Falconer, S.M., Coulet, A., Noy,

N.F., Musen, M.A., & Shah, N.H., 2011. NCBO

resource index: ontology-based search and mining of

biomedical resources, Journal of Web Semantics, vol.

9, no. 3, pp. 316–324.

Kuo, T.-T., Tseng, S.-S. & Lin, Y.-T., 2003. Ontology-

based knowledge fusion framework using graph

partitioning. In: P.W.H. Chung, C.J. Hinde, M. Ali

(eds.), IEA/AIE 2003, 16th international conference on

industrial and engineering applications of artificial

intelligence and expert systems, Laughborough, UK,

23–26 June 2003. Lecture notes in artificial

intelligence, vol 2718, pp. 11–20.

Lee, K.R., 2007. Patterns and processes of contemporary

technology fusion: the case of intelligent robots. Asian

Journal of Technology Innovation, vol. 15, no. 2, pp.

45–65.

Lin, L.Y. & Lo, Y.J., 2010. Knowledge creation and

cooperation between cross-nation R&D institutes,

International Journal of Electronic Business

Management, vol. 8, no. 1, pp. 9–19.

Nengfu, X., Wensheng, W., Xiaorong, Y. & Lihua, J.,

2012. Rule-based agricultural knowledge fusion in

web information integration, Sensor Letters, vol. 10,

no. 8, pp. 635–638.

Liang, P., Ling, H., Blasch, E., Seetharaman, G., Shen, D.,

Chen, G. 2013. Vehicle detection in wide area aerial

surveillance using Temporal Context. In: Proceedings

of the 16th international conference on information

fusion, Istanbul, Turkey, 9–12 July 2013, pp. 181–188.

Preece, A., Hui, K., Gray, A., Marti, P., Bench-Capon, T.,

Cui, Z. & Jones, D., 2001. Kraft: an agent architecture

for knowledge fusion, International Journal of

Cooperative Information Systems, vol. 10, no. 1–2, pp.

171–195.

Roemer, M.J., Kacprzynski, G.J. & Orsagh, R.F., 2001.

Assessment of data and knowledge fusion strategies

for prognostics and health management. In: 2001 IEEE

Aerospace conference, Big Sky, Montana, USA, 10–

17 March 2001, vol 6, pp. 2979–2988.

Sanchez-Pi, N., Martí, L., Molina, J.M., Garcia, A.C.B.,

2014. High-Level Information Fusion for Risk and

Accidents Prevention in Pervasive Oil Industry

Environments. In: Highlights of Practical Applications

of Heterogeneous Multi-Agent Systems. The PAAMS

Collection, vol. 430, pp. 202–213.

Smirnov, A., Kashevnik, A., Shilov, N., Balandin, S.,

Oliver, I. & Boldyrev, S., 2010. On-the-fly ontology

matching in smart spaces: a multi-model approach. In:

Smart Spaces and Next Generation Wired/Wireless

Networking, Proceedings of the third conference on

smart spaces, ruSMART 2010, and the 10

international conference NEW2AN 2010, St.

Petersburg, Russia, 23–25 August 2010. Lecture notes

in computer science, vol. 6294, pp. 72–83.

Smirnov, A., Levashova, T. & Shilov N., 2013. Patterns

for context-based knowledge fusion in decision

support, Information Fusion. http://dx.doi.org/10.1016

/j.inffus.2013.10.010 (in press).

Smirnov, A., Pashkin, M., Chilov, N. & Levashova, T.,

2005a. Constraint-driven methodology for context-

based decision support, Journal of Decision Systems,

vol. 14, no. 3, pp. 279–301.

Smirnov, A., Pashkin, M., Chilov, N., Levashova, T. &

Haritatos, F., 2003. Knowledge source network

configuration approach to knowledge logistics,

International Journal of General Systems, vol. 32, no.

3, pp. 251–269.

KnowledgeFusioninContext-AwareDecisionSupportSystems

193

Smirnov, A., Pashkin, M., Levashova, T. & Chilov, N.,

2005b. Fusion-based knowledge logistics for

intelligent decision support in network-centric

environment. International Journal of General

Systems, vol. 34, no. 6, pp. 673–690.

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

194