1212
5, MAY 2013
Ontology-based Semantic Retrieval for Risk
Management of Construction Project
Shaohua Jiang
Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian, China
Email: shjiang@dlut.edu.cn
Jian Zhang, and Haiyan Zhang
Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian, China
Email: zhangjian_nico@hotmail.com, molihaiyan@126.com
AbstractRisk management is crucial for successful
construction project management. Effective risk assessment
can improve construction enterprises competitiveness and
profitability to a great extent. And the information of
previous projects is a very precious resource and reference
for construction project risk management. In order to
facilitate project participants to query required information
of previous projects among numerous project documents
efficiently, this paper introduces ontology as the modeling
approach for information exchange and puts forward a
semantic retrieval method based on this ontology for
construction project risk management. At first, this paper
introduces the procedure of constructing risk domain
ontology by means of seven footwork modeling method, then
the methods of construction projects document processing
and querying sentences expansion are presented. Finally,
this paper put forward a semantic retrieval system based on
construction project risk domain ontology which can
correlate the ontology conceptual knowledge and the
documents content. Experimental results suggest that the
proposed ontology-based semantic retrieval method can
improve informations semantic expression ability and
realize the intellectualization of construction project risk
management.
Index Termsconstruction project; risk domain ontology;
document processing; query expansion; semantic retrieval
I. INTRODUCTION
Risk management is a formal process of identifying,
analyzing and responding to risk events systematically
throughout the lifecycle of a project to obtain an
acceptable degree of risk elimination or control [1]. Due
to the indefinite and ambiguous external and internal
environment of construction project, risk management is
a difficult task. Although every project is unique,
experiences of past projects still play a significant role in
managing new projects. These experiences can accelerate
the construction process and minimize construction cost
through experiences reusing and sharing among
engineers and participants. A couple of reference
frameworks and models have been developed to provide
systemic methods for construction project risk
management [2, 3]. Its obvious that the successful
2013 ACADEMY PUBLISHER
doi:10.4304/jnw.8.5.1212-1220
JOURNAL OF NETWORKS, VOL. 8, NO.
applications of these frameworks depend on the risk-
related information (such as likelihood of risk, potential
impact, risk allocation between the parties etc.) fed into
the frameworks. The successful applications of risk
models depend on the identified risk factors, the relation
between these risk factors and the risk propagation mode.
One of the main drawbacks of these reference
frameworks and risk models is lacking a universal
vocabulary. Learning from past projects can contribute
significantly to the creation of realistic risk scenarios and
the development of reliable risk model so as to manage
risks successfully. Thus this paper introduces ontology as
the modeling approach for risk-related information
exchange of construction project.
The concept of ontology stems from philosophy
domain and focuses on the abstract expression of
objective reality. Ontologys application in other fields
derives from Artificial Intelligence (AI) domain. Among
various definitions of ontology, Gruber (1993) [4] from
Stanford University put forward the definition of
ontology in the field of AI, namely "Ontology is an
explicit specification of conceptualization", which is
quoted most. Later Borst Pim (1997) [5] supplemented
the definition from his point of view, "Ontology is a
formal specification of a shared conceptualization".
Studer etc. (1998) [6] made a further research on the
above two definitions, and in his opinion, ontology is a
formal, explicit specification of a shared
conceptualization and contains four key features:
conceptualization, explication, formalization and share[7].
The purpose of utilizing ontology is to access relevant
domain knowledge, provide common understanding of
domain knowledge, confirm common recognized terms of
a domain and provide clear definition of the terms and
relationship between these terms from different levels of
the formal model [8], [9]. In recent years, the importance
of ontology in cooperative work has been realized widely,
and research related to ontology and semantic retrieval
increases rapidly.
F. Staub-French et al. (2003) [10] presented an
ontology which was constructed for relating building
product models to construction activities and related
construction resources in order to calculate construction
JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013
cost. Compared with traditional tools, this proposed
ontology enabled estimators to make construction cost
estimates more completely and consistently. O. Ugwu et
al. (2005) [11] developed an ontological model for
constructability assessment of steel structures. C. Lima et
al. (2005) [12] presented e-Cognos project, and it is an
ontology-based portal for construction knowledge
management. The study of A. Aksamija and F. Grobler
(2007) [13] focused on developing an ontology which
described the principles of building design and the
relationships between them. Q. Li (2011) [14] introduced
an ontology to construction project document
management process, constructed a contract domain
ontology of construction project and analyzed the
interaction between BIM model object data and the
ontology-based database. H. Wang et al. (2011) [15]
constructed an ontology for civil aviation emergency
decision and implemented an ontology based rules
reasoning by using protg and Jena, his research
provided method and technology for the knowledge
management and application of civil aviation emergency
decision. X. L. Liu (2012) [16] constructed an education
domain ontology in his masters thesis and discussed the
semantic retrieval method through researching and
assessing the constructed ontology. L. Yang (2012) [17]
constructed a paddy rice domain ontology in his masters
thesis and realized the ontology based information
retrieval system so as to improve the precision of word
retrieval in paddy rice domain.
These research laid foundation for ontology based
semantic retrieval, but its not enough for risk
management of construction project. Due to the lack of
universal vocabulary and invalidity of retrieving
information from concept level, therefore, based on the
Main Process
Specification
determination of the scope
Conceptualization
the collection and organization of the
relevant domain concepts
Formalization
the representation of knowledge in a
formal way
Implementation
converting the knowledge into a
machine-comprehensible language
Evaluation
validating the completeness and the
generality of the ontology
Figure 1. Ontology development process
2013 ACADEMY PUBLISHER
1213
above research, this paper proposes an ontology based
semantic retrieval system to facilitate construction project
risk management.
Firstly, this paper constructs an ontology which can be
used as the foundation of construction projects risk
events database. This ontology includes risk sources, risk
events and risk consequences. The purpose for
constructing risk domain ontology is to provide a unified
risk-related concept set through illustrating concepts and
the relationships between them. This ontology can be
utilized, shared and reused by people and computers and
lay a solid foundation for semantic retrieval. Then this
paper introduces methods for construction projects
document processing and querying sentences expansion.
Finally, this paper develops a semantic retrieval system
based on construction project risk domain ontology so as
to make users query about risk-related questions more
targeted and improve the efficiency of construction
project risk management in a great extent.
II. DEVELOPMENT OF RISK DOMAIN ONTOLOGY FOR
CONSTRUCTION PROJECT
Ontology development includes five main steps:
standardization (determination of scope),
conceptualization (collection and organization of the
relevant domain concepts to be included in ontology),
formalization (knowledge representation in a formal way),
implementation (converting the formalized knowledge
into a machine-comprehensible ontology language) and
evaluation (validating the completeness and the
generality of the ontology). Fig. 1 demonstrates the
process of ontology development and the activities
supported by each step [18].
Output
identifying the limits and requirements of
risk management in the current project
determining the important concepts, their
properties and relationships between them
developing the formal representation of
relationships between concepts
protg modeling
interviewing with experts
1214 JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013

Due to the diversity of risk results, such as low-quality,
cost overrun and lower customer satisfaction, its a long-
term process of accumulation to construct an ontology
which can support all risk management tasks for
construction project. This paper only constructs an
ontology aiming at the risk of cost overrun. The related
knowledge of construction projects risks about cost
overrun is presented by detailed literature review, so the
standardization stage is achieved and an original model
containing several concepts is developed. Then relevant
experts of construction project risk management are
asked to give suggestions about these concepts and the
mutual relationship among them, and some concepts are
restated and regrouped based on experts suggestions. For
example, it is found that a countrys inflation rate, tax
rate and exchange rate play important roles on economic
environment by summarizing related literatures. However,
through communicating with experts, it is realized that
the stability of them is much more important than the
aforementioned values. Therefore, these parameters are
grouped in the same title adverse change.
In conceptualization phase, in order to improve data
structure, various concepts are organized into a
hierarchical form, which is called classification criterion.
In one model, classification criterion presents the
classification of elements and lays a foundation for reuse
and tasks integration. This stage is one of the most
difficult activities in ontology design process, which not
only includes the objective presentation of the world, but
also describes the views of people to the world and the
classification of these things. Fig. 2 shows a hierarchical
graph of ontology for construction project risk
management about cost overrun [2].

Risk Concepts

Risk
Risk Sources Risk Events
Consequences

Unexpected Adverse Cost Overrun

Events Changes (AC)

Unexpected Wars and
AC in Weather Situation
Riots
Unexpected Natural
Disasters
Unexpected Accidents
Unexpected Labor
Problems
Unexpected Social
Problems
Unexpected Historical
Problems
Decreasing Productivity
AC in Tax Rate
Decreasing Construction
Quality
AC in Geography AC in Exchange Rate
Increasing Work Quantity
AC in Relations with AC in Policies and
Government Regulations
Increasing Unit Costs of
Resources
AC in Relations with
AC in Project Scope
Cooperators
Delaying Work Progress
AC in Relations with
AC in Project Design
Designers
Logistics Delay
AC in Relations with
AC in Project Teams
Engineers

AC in Relations with AC in Availability of

Clients Labors

AC in communication AC in Availability of
between Parties Materials

AC in
AC in Availability of
ContractorsEconomic
Equipments
Conditions
AC in
AC in Availability of Sub-
CooperatorsEconomic
contractors
Conditions

AC in Construction
AC in ClientsAttitudes
Technology

AC in Original Schedule AC in Public Reaction

Figure 2. Hierarchical graph about construction project risk field on cost overruns

2013 ACADEMY PUBLISHER

JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013
Figure 3. Ontology of construction project risk domain
Figure 4. Partial OWL source code of risk domain ontology
In the formalization stage, a complete ontology which
is suitable for current circumstances is generated step by
step through using a progressive development method. A
standard ontology is completed eventually by clarifying
concepts, concepts attributes, the restricted condition of
these attributes and the relationship between these
concepts. Formalized knowledge is transformed into
framework based representation of protg. Protg is a
free and open platform developed by Stanford Medical
Informatics and Knowledge System Library in Java
environment. Compared with other systems, protg has
some obvious advantages, such as graphical user
interface, supporting input and output format of RDF,
UML, OWL etc. and free downloading etc. [19]. OWL
refers to web ontology language, which is developed by
World Wide Web Consortium (W3C), and this language
is used for describing ontology. Protg 4.2 is adopted to
construct risk domain ontology in this paper. Fig. 3 and
2013 ACADEMY PUBLISHER
1215
Fig. 4 show the risk domain ontology and the partial
OWL source code of this ontology respectively.
In the implementation stage, classes, their slots, the
facets, instances and relationship between classes become
part of the software presented.
Finally, in the evaluation stage, the completeness level,
the universality level and the effectiveness level are
tested to validate the developed ontology. These
measurement standards are confirmed at last through
interactive seminar and interview with experts.
III. SEMANTIC RETRIEVAL BASED ON CONSTRUCTION
PROJECT RISK DOMAIN ONTOLOGY
To implement semantic retrieval based on construction
project risk domain ontology, several sequential steps, i.e.,
segmentation, eigenvalue extraction and weight
calculation, query expansion, similarity calculation, need
to be finished.
1216
A. Segmentation Processing
In this section, related construction document is
processed by text segmentation. Currently, there are some
word processing softwares. ICLCLAS [20] is adopted to
deal with the documents in this paper. ICTCLAS is
developed by Chinese Academy of Sciences, and it
supports identifying new words, marking the part of
speech, and using user dictionary, etc. As ICLCLAS
handles word segmentation according to the common
vocabulary, the proprietary vocabulary in the domain of
construction project risk management cannot be
accurately extracted. So this paper inputes the constructed
construction project risk domain ontology into the
common vocabulary in order to achieve a more accurate
word segmentation processing result.
B. Eigenvalue Extraction and Weight Calculation
After related construction document is processed by
ICLCLAS, the adverbs, conjunctions, adjectives and
other parts of speech of words which can't express the
document content very well are filtered out, and the rest
are mainly the verbs and the nouns. Next, this paper uses
the TF-IDF [21] algorithm to calculate the document
frequency of the verbs and the nouns. The more frequent
a word is appeared in a document, the more relevant it is
to the topic of the document. In this process, w(d, t) is
dened as the weight of a term t in document d ,
N represents the total number of documents, and n
illustrates the number of documents which contain the
terms t . And the standard TF-IDF formula and IDF
formula are as follows [21]:
N
tf(d, f)lg( )
nt
wd, t=
m
N Assuming query vector is X and eigenvector of
tf (d, t j ) (lg( )) 2
2
n construction project document is Y , the similarity
j=1 t
N [21]:
IDF = lg
nt
Weight of the terms in document can be calculated by
this formula. As any document covers many terms, a
threshold can be set according to the need. The term
whose weight is lower than the threshold can be filtered
out. Its assumed that a query is composed of two terms, t1
and t2 , and the normalized weights associated with t1 and t2
are wt d and wt d respectively. Then the document can be
1 2
expressed as vector ( wt d , wt d ) .
1 2
C. Query Expansion
Query expansion refers to expand the user's query term
to form a longer query vector to obtain more accurate
query result. In the process of querying the questions
about construction project risk, its a common problem
that query terms are too short so as to lead to the
ambiguity of query problem and more redundant
documents which affect query efficiency. This paper
2013 ACADEMY PUBLISHER
JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013
introduces logical reasoning and ontology knowledge to
expand query terms. If query term is a concept of the
constructed ontology, the methods of synonym expansion
and generalization extension can be used to expand query,
then the semantic similarity of obtained terms and query
terms can be calculated, and the terms with higher
similarity are chosen as the last extension terms that
constitute the query vector [22].
If the query term is not a concept of the ontology, the
expansion term can be extracted according to the context
analysis method presented by D. Fensel etc. (1998) [23].
Generally, if the context situation of two terms is similar,
these two terms are related with each other to some extent.
Therefore, in case the context situation of a certain
concept in the ontology and query term is similar, the
concept must have close relationship with the query term,
and can be used as the final expansion term. To sum up,
the final query vector consists of the query term and the
expansion term.
D. Similarity Calculation
There are two methods, i.e., distance formula algorithm
and cosine formula algorithm, which can be used to
calculate the similarity between query vector and text
vector.
Let N represents the collection of all eigenvectors, pi
and qi represent the vector component of text vector p
and q in the dimension of i , d represents the distance
between the query vector and text vector, the distance
formula for similarity calculation is as follows [21]:
N
d= (p
i =1
i - qi )2 (3)
(1)
between these two vectors can be calculated as follows
(2)
cos =
(X Y ) (4)
X Y
The similarity between users query vector and the
construction documents eigenvector can be calculated
accurately by this formula. Documents with lower
similarity can be filtered out by predefined similarity
threshold, and the most suitable documents can be
retrieved and the retrieval efficiency can be improved.
Fig. 5 shows the risk domain ontology based retrieval
system for construction project risk management.
IV. EXPERIMENTAL RESULTS AND ANALYSIS
To verify the developed ontology-based semantic
retrieval method for risk management of construction
project, query expansion and semantic similarity
calculation and ranking experiments are conducted in this
paper.
JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013 1217

Construction
UsersQuery Document
Stament Database

Segmentation
Segmentation Processing
Processing
Risk Domain
Ontology Eigenvalue
Extraction

Query
Expansion
Weight
Calculation

Document
QueryVector
Vector

Similarity
Calculation

Users

Figure 5. Risk domain ontology based retrieval system framework

based on the definition and formulas (5). Then the

A. Query Expansion
resulted terms, C2 C6, can be ranked based on the
Partial risk domain ontology is shown as Figure 6. similarity extent and the terms with bigger semantic
Query expansion is studied and realized based on this similarity are selected as the expansion terms [24].
ontology under the condition of the query terms are To explain the formula (5) clearly, two definitions are
concepts of the ontology. given as follow.
Definition 1. For two concepts C1 and C2 in a same
ontology, if C1 C2, then C1 is the concept super of C2;
Risk Concepts

if C1 C2, then C1 is the l concept lower of C2.

Definition 2. The concept super set ( CS ) of concept C
Risk
Risk Sources Risk Events
Consequences
includes concept C and all its concept supers. Because CS
contains concept C at least, CS . The semantic
similarity of two concepts Ci and C j in one ontology O
can be represented as ( Ci , C j ) [ 0,1] as formula (5),
Decreasing
Unexpected Adverse Decreasing
Construction Cost Overrun
Events Changes (AC) Productivity
Quality
Ci , C j O
Increasing Unit
Increasing
Costs of
Work Quantity
Resources 1 Ci S C j S


Ci S C j S
Delaying Work Logistics (C i , C j ) = Ci S C j S O (5)
Progress Delay
Ci S C j S


0 Ci S C j S =
Figure 6. Partial risk domain ontology
The following results can be calculated by the above
Let the query statement is represented as follows: formula.
(C 2, C1) =
1
Q: unexpected events (C1). ,
3
Because the query term, unexpected events, is one of
2
the concepts in risk ontology, query can be generalized to (C 3, C1) = ,
3
expansion by the constructed ontology. The resulted 1
terms after generalization are risk concepts (C2), risk (C 4, C1) = ,
4
source (C3), risk event (C4), decreasing 1
productivity (C5) and adverse change (C6). (C 5, C1) = ,
5
The semantic similarity can be calculated between the 1
query term, C1, and the terms, C2 C6, respectively (C 6, C1) =
2

2013 ACADEMY PUBLISHER

1218 JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013

So the order of semantic similarity between term in
ontology and query terms is as follows:
(C 3, C1) > (C 6, C1) > (C 2, C1)
> (C 4, C1) > (C 5, C1)
So the terms with the biggest and smallest semantic
similarity to unexpected events are risk source and
decreasing productivity respectively. If the predefined
threshold is 0.5, the three terms, risk event, risk
concepts, decreasing productivity, can be filtered out
because their semantic similarities are lower than 0.5. At
last, risk source and adverse change are selected as
query expansion terms to construct query vector.
B. Semantic Similarity Calculation and Ranking
According to the expansion result of query term
unexpected events in the previous section, risk
source and adverse change are chosen as the
expansion terms which form the query vector. Then the
similarity between query vector and text vector can be
calculated before and after the query expansion, and the
similarity order can be determined. The query statement,
Q1 and Q2, and documents, D1, D2, D3, D4, D5, D6 and
D7 are as follows:
Q1: unexpected events
Q2: unexpected events, risk source, adverse
changes
D1: Both unexpected events and adverse changes are
the risk source of a project.
D2: Unexpected events would greatly affect a
construction projects cost.
D3: In the construction process of a project, there are
lots of unexpected risks.
D4: The risk sources of a construction project can
result in risk events in some cases.
D5: During the construction phase of a project, many
unexpected changes can occur.
D6: In the lifecycle of a project, unexpected wars and
riots are part of the unexpected events that can affect the
cost of a project.
D7: Adverse changes in tax rate can cause some risk
events during a projects lifecycle.
Table I. shows the IDF value and term weight of every
eigenvalue of seven documents.

TABLE I.
IDF VALUE AND TERM WEIGHT OF EVERY EIGENVALUE OF THREE DOCUMENTS
Term weights
T\D IDF D1 D2 D3 D4 D5 D6 D7 Q1 Q2
Adverse 0.544 0.544 0 0 0 0 0 0.544 0 0.544
Affect 0.544 0 0.544 0 0 0 0.544 0 0 0
Case 0.845 0 0 0 0.845 0 0 0 0 0
Cause 0.845 0 0 0 0 0 0 0.845 0 0
Change 0.368 0.368 0 0 0 0.368 0 0.368 0 0.368
Construction 0.243 0 0.243 0.243 0.243 0.243 0 0 0 0
Cost 0.544 0 0.544 0 0 0 0.544 0 0 0
Events 0.146 0.146 0.146 0 0.146 0 0.146 0.146 0.146 0.146
Lifecycle 0.544 0 0 0 0 0 0.544 0.544 0 0
Occur 0.845 0 0 0 0 0.845 0 0 0 0
Part 0.845 0 0 0 0 0 0.845 0 0 0
Phase 0.845 0 0 0 0 0.845 0 0 0 0
Process 0.845 0 0 0.845 0 0 0 0 0 0
Project 0 0 0 0 0 0 0 0 0 0
Rate 0.845 0 0 0 0 0 0 0.845 0 0
Result 0.845 0 0 0 0.845 0 0 0 0 0
Riots 0.845 0 0 0 0 0 0.845 0 0 0
Risk 0.243 0.243 0 0.243 0.486 0 0 0.243 0 0.243
Source 0.544 0.544 0 0 0.544 0 0 0 0 0.544
Tax 0.845 0 0 0 0 0 0 0.845 0 0
Unexpected 0.146 0.146 0.146 0.146 0 0.146 0.292 0 0.146 0.146
Wars 0.845 0 0 0 0 0 0.845 0 0 0

The similarity calculation process is as follows:
Cos (Q 1 , D 1 ) = 0.1460.146+0.1460.146=0.042
Cos (Q 1 , D 2 ) = 0.1460.146+0.1460.146=0.042
Cos (Q 1 , D 3 ) = 0.1460.146=0.021
Cos (Q 1 , D 4 ) = 0.1460.146=0.021
Cos (Q 1 , D 5 ) = 0.1460.146=0.021
Cos (Q 1 , D 6 ) = 0.1460.146+0.1460.146=0.042
Cos (Q 1 , D 7 ) = 0.1460.146=0.021
Cos (Q 2 , D 1 )
=0.5440.544+0.3680.368+0.1460.146+0.2430.24
3+0.5440.544+0.1460.146=0.829
Cos (Q 2 , D 2 ) =0.1460.146+0.1460.146=0.043
Cos (Q 2 , D 3 ) = 0.2430.243+0.1460.146=0.080
Cos (Q 2 , D 4 )
=0.1460.146+0.2430.486+0.5440.544=0.435
Cos (Q 2 , D 5 ) =0.3680.368+0.1460.146=0.157
Cos (Q 2 , D 6 ) =0.1460.146+0.2930.146=0.064
Cos (Q 2 , D 7 )
=0.5440.544+0.3680.368+0.1460.146+0.2430.24
3=0.509
The similarity calculation results between the query
vectors and document vectors are shown in Table II.

2013 ACADEMY PUBLISHER

JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013
TABLE II.
THE COMPARISON RESULTS BETWEEN QUERY VECTORS AND DOCUMENT VECTORS
D
Similarity D1 D2
Similarity(Q1) 0.042 0.042
Similarity(Q2) 0.829 0.043
According to the similarity calculation results, the
similarity rankings are as follows:
When the query vector is Q1, the result is:
D 1 =D 2 =D 6 >D 3 =D 4 =D 5 =D 7
While the query vector is Q2, the result is:
D 1 >D 7 >D 4 >D 5 >D 3 >D 6 >D 2
It can be concluded from the above results that when
unexpected events is used as the only query term,
document D1, D2 and D6 are selected as relevant
documents. While after the query term is expanded,
document D2 and D6 with lower similarity are filtered
out and only document D1 is selected as the final
reference information to meet the requirement of user
better. So its verified that the semantic retrieval method
based on risk domain ontology can improve the retrieval
accuracy and efficiency for construction project risk
management.
IV. CONCLUSION
Risk management is crucial for successful construction
project management. This paper provides an ontology-
based semantic retrieval method to facilitate utilizing
previous projects experience for risk management of
construction project. Firstly, this paper constructs an
ontology for risk management of construction project.
This ontology provides a general vocabulary of project
risk domain. The vocabulary forms a main frame for risk
management information model, and constructs the basic
structure of risk event history database. Then on the basis
of the risk domain ontology, this paper introduces the
semantic retrieval method and proposes a systematic
method based on ontology semantic retrieval for
construction project risk management. Finally,
experimental results verify that this method can improve
the accuracy and efficiency of users retrieval to a
significant extent, thus realize high efficient usage of past
project experience for project participants to improve
construction project risk management. However, the
scope of this construction project risk domain ontology is
only focused on the risks of cost overrun, and other
possible project objectives, such as quality, duration, and
customer satisfaction etc., are not covered by the
ontology. In the following study, the authors intend to
study domain ontology related to other risks, and support
the construction project risk management more
comprehensively.
ACKNOWLEDGMENT
2013 ACADEMY PUBLISHER
1219
D3 D4 D5 D6 D7
0.021 0.021 0.021 0.042 0.021
0.080 0.435 0.157 0.064 0.509
This work was supported by National Natural Science
Foundation of China (Grant No. 51178084).
REFERENCES
[1] J. AlBahar and K. Crandall, Systematic risk
management approach for construction projects, Journal of
Construction Engineering and Management, vol. 116, pp.
533546, September 1990. doi:org/10.1061/(ASCE)0733-
9364(1990)116:3(533)
[2] G. Fidan, I. Dikmen, A. M. Tanyer and M. T. Birgonul, An
ontology for relating risk and vulnerability Journal of
Computing in Civil Engineering, vol. 25, pp. 23-73, August
2010. doi:10.1061//(ASCE)CP.1943-5487.0000090
[3] P. Liu, C. Xu, X. X. Wang, X. Y. Wang and G. H. Xu, A
statistical amalgamation approach for ontologies, Journal
of Networks, vol. 7, pp. 243249, Febuary 2012.
doi:10.4304/jnw.7.2.243-248
[4] T. R. Gruber, A translation approach to portable ontology
specification, Knowledge Systems Laboratory Technical
Report, pp. 199220, 1993.
[5] P. Borst and H. Akkermans, An ontology approach to
product disassembly, Sant Feliu de Gu5xols, pp. 3348,
1997.
[6] R. Studer, V. R. Benjamins and D. Fensel, Knowledge
engineering, principles and methods, Data and knowledge
engineering, vol. 25, pp. 161197, March
1998. doi:org/10.1016/S0169-023X(97)00056-6
[7] C. S. Lee, Y. F. Cao, Y. H. Kuo and M. H. Wang,
Automated ontology construction for unstructured text
documents, Data & Knowledge Engineering, vol. 60, pp.
547566, March 2007. doi:org/10.1016/j.datak.2006.04.001
[8] W. Gao and T. W. Xu, Characteristics of optimal function
for ontology similarity measure via multi-deviding,
Journal of Networks, vol. 7, pp. 12511259, August
2012. doi:10.4304/jnw.7.8.1251-1259
[9] M. P. Nepal, S. Staub, R. Pottinger and J. Zhang,
Ontology-based feature modeling for construction
information extraction from a building information model,
Journal of Computing in Civil Engineering, vol. 23, pp. 13
64, July 2012. doi:org/10.1061/(ASCE)CP.1943-
5487.0000230
[10] S. Staub, M. Fischer, J. Kunz and B. Paulson, An
ontology for relating features with activities to calculate
costs, Journal of Computing in Civil Engineering, vol. 17,
pp. 243254, February 2003. doi:org/10.1061/(ASCE)0887-
3801(2003)17:4(243)
[11] O. Ugwu, C. L. Anumba and A. Thorpe, Ontological
foundations for agent support in constructability assessment
of steel structures, Automation in Construction, vol. 14, pp.
99114, October
2008. doi:org/10.1016/j.autcon.2004.08.009
[12] Ontology-based Optimization of Knowledge
Management in Construction, e-Construction. [Online].
Available from: http://www.itcon.org/2005/21/
1220
[13] A. Aksamija and F. Grobler, Architectural ontology:
development of machine-readable representations for
building design drivers, International Workshop on
Computing in Civil Engineering, pp. 168175, 2007.
[14] Q. Li, Research on ontology-based document
management system under BIM environment, Dalian
University of Technology, Dalian, 2011.
[15] H. Wang, X. Yang, J Wang and L. Li, Research on
ontology based knowledge presentation and reasoning in
civil aviation emergency decision, Computer Engineering
& Science, vol. 33, pp. 129133, April 2011.
doi:1007130X( 2011)04012905
[16] X. L. Liu, The construction and application research
of the top-level ontology of educational domain, Jilin
University, Changchun, 2012.
[17] L. Yang, The research of semantic retrieval based on
domain ontology, Anhui University, Anhui, 2012.
[18] M. Fernandz-Lopz, A. Gomz-Perz and N. Juristo, information and knowledge management .
Methontology: from ontological arts towards ontological
engineering, Proceedings of the AAAI97 Spring
Symposium Series on Ontological Engineering, pp. 3340,
1997.
[19] S. H. Jiang, J. Zhang and Q. Li, Study on contract
domain ontology supporting BIM interactivity, 2012
International Conference on Construction and Real Estate
Management (ICCREM 2012), pp. 420423, 2012.
[20] K. Y. Lin, L. Soibelman, Promoting transactions for
A/E/C product information, Automation in Construction,
vol. 15, pp. 746757, November
2006. doi:org/10.1016/j.autcon.2005.09.008
[21] G. Lee, C.M. Eastman, R. Sacks and S.B. Navathe,
Grammatical rules for specifying information for
automated product data modeling, Advanced Engineering
Informatics, vol. 20, pp. 155170, April
2006. doi:org/10.1016/j.aei.2005.08.003
[22] S. H. Liu, Y. Gao, M. Li, P. Kilaru, T. Smith and S.
Toner, A semantics and data-driven biomedical multimedia
software system, Journal of Multimedia, vol. 5, pp. 353
362, August 2010. doi:10.4304/jmm.5.4.352-360
[23] D. Fensel, S. Docker, M. Edrmann and R. Studer,
How to enable intelligent access to WWW, Proceedings
2013 ACADEMY PUBLISHER
JOURNAL OF NETWORKS, VOL. 8, NO. 5, MAY 2013
of the 11th Banff Knowledge Acquisition for Knowledge-
Based System Workshop, pp. 117, 1998.
[24] N. Y. Shi and C. Yang, Towards domain ontology-
based semantic annotation research, Computer
Engineering and Design, vol. 28, pp. 59855987, December
2007. doi:10.3969/j.issn.1000-7024.2007.24.049
Shaohua Jiang was born in Wafangdian, Liaoning Province,
China in 1971. He received his Ph.D degree in Management
Science and Engineering from Dalian University of Technology,
Dalian, China in 2006. His major field of study is construction
management.
He is an Assistant Professor in faculty of Infrastructure
Engineering at Dalian University of Technology, China. His
research interests include wireless sensor network, building
Jian Zhang was born in Siping, Jilin Province, China in 1987.
She received her bachelor degree in Construction Management
from Jilin University, Changchun, China in 2010. Her major
field of study is construction management.
She is currently a postgraduate student in faculty of
Infrastructure Engineering at Dalian University of Technology,
China. Her research interests include building information
model, information management.
Haiyan Zhang was born in Nantong, Jiangsu Province, China
in 1987. She received her bachelor degree in Construction
Management from Jiangnan University, Wuxi, China in 2010.
Her major field of study is construction management.
She is currently a postgraduate student in faculty of
Infrastructure Engineering at Dalian University of Technology,
China. Her research interests include text mining, building
information model, information management.