context based ontology inference system using multi ... lee.pdf · decision making criteria, we...
TRANSCRIPT
1
Context Based Ontology Inference System Using Multi-Criteria
Decision Making
JaeGil.Lee*1, YongJin.Joo2, SooHong.Park3
([email protected], [email protected], [email protected])
Abstract
LBS services are user-centric location based information services,
whose importance has been discussed as an essential engine in
an Ubiquitous Age. Recent researches have provided intelligent
services which provide relevant information on the basis of
information on users through context awareness, instead of
providing information merely. However, current researches
provide information to users through single-reference reasoning
using basic information on users and preference information, but
there is a limit that they couldn't reflect decision making according
to users' preference and their selection standards. So, this study
has constructed an ontology of basic information on users,
preference and space preference and cost value ontology of
giving weights to selection standards for decision-making, in order
to overcome such limit. Accordingly, it suggests an ontology
reasoning system that enables users to derive recommended
results suitable for them through selection standard reasoning
according to users' various preferences. In case of construction of
the ontology, it could construct personal characteristics of users,
knowledge on preference and knowledge on spatial and
geographical preference and integrate a function of reasoning
relevant information through construction of cost value ontology
1 Dept. of GeoInformatic Engineering, In-ha University, Master`s course 2 Institute of Urbarn Science, University of Seoul, Research Professor 3 Dept. of GeoInformatic Engineering, In-ha University, Associate Professor
2
for various selection standards with functions of spatial data
management of GIS system, spatial query and map display. Thus,
this study showed it could support multi-criteria decision making
by giving weight according to users' preference using AHP
technique, instead of service provision through simple context
awareness of users, and provide actual and personalized
information to users.
Key words: Spatial information, intelligent geospatial service,
service classification matrix, service domain, service intelligent
level, geo-location accuracy
1. INTRODUCTION
1.1 Research Background & Objectives
The context awareness means the recognition of individual’s location, time, weather,
schedule and personal preference. Current location based services(LBS) provides
the information and services simply with the users’ context based on the users’
current location including map, weather, traffic situation. .
Picture 1) Context-aware
Semantic technology was recently adapted in the LBS area while trying to get out of
providing simple information. Semantic technology is an intellectual technology
which understands the meaning in the system by giving meanings to the information.
3
In order to apply the semantic technology we need the formation of systematic
mechanism for knowledge interface and the concept of the ontology. Using ontology,
it is possible to support the decision making through the communication between
users and system, and provide intelligent and flexible services which recommend the
most suitable services to the users.
Recent researches provide users with the information through the interface of single
criteria using information of users and preference. This has a limit which can not
reflect the users’ decision making with various criteria.
In order to overcome the limit, we architected ontology according to location, basic
information, preference and space preference by interpreting and context awareness
of users. After architecting the cost value ontology to give weight on the factors of
the decision making criteria, we suggest the service measures based on location can
provide the users with the recommended suitable result through multi-criteria
decision interface.
1.2 Research Configuration
In this research, we architected ontology according to location, basic information,
preference and space preference by interpreting and context awareness of users.
After architecting the cost value ontology to give weights on the elements of the
decision making criteria, we architect a GIS interface system which shows the
suitable recommended results for the users’ current location and context through
multi-criteria decision interface.
First, we analyze the previous context aware – location based systems, and investigate
the possibility of unification with the geographic data system after studying previous
researches which are based on the theories related to ontology.
Second, in order to architect ontology, we define using the Ontology Web Language
(OWL) which is the standard of the World Wide Web Consortium (W3C).
Third, since the experimental GIS interface system has spatial question functions, we
architect the ontology for users’ context awareness, individual characteristics, space
preference and the search results come from the corresponding spatial questions. And
4
then, we architect the cost value ontology which is to consider the weight of the factors for
various decision making criteria.
Fourth, we apply the GIS inference system which is embodied in the area for the
experiment. After analyzing the actual interfaced result, we seek the actual
application possibility of GIS inference system through using ontology.
We used several programs in order to perform the research. To architect the
ontology we used Protege4.1-alpha, Google Map API for spatial database and the
system was developed by Apple iOS 4.01.
This research is described like below. In chapter 2, we explain the technology and the
research related to the concept of ontology. In chapter 3, we architect the ontology
according to the users’ individual characteristics and the space preference through the
users’ context awareness, and then explain about the materialization of the GIS interface
system. In chapter 4, we show the conclusion of this research and the direction of the
researches in the future.
2. Literature Review
2.1 Previous context awareness LBS
Most of the LBS have provided information based on the database. Referring
various information and search patterns according to the users and location
information, database is designed.
After inputting the information suitable for the criteria, the information match to the
condition is provided through query. In this kind of database based search method,
there could be difficulties in search about complex questions related to users’
different context because this method designs database and saves the information
according to the forecasted criteria.
In order to provide more intelligent LBS, studies of context aware LBS which
provide LBS matches for the users’ demand and considers user’s situations are
recently introduced.
After architecting ontology with consideration for users’ location information,
schedule, information of service, service operating hours and costs, these
5
researches provide users with suitable services through interface for information
matches the meaning.
However, these researches did not consider users’ preference and space
preference, and cannot support analytic decision making according to the
importance of various factors for decision making.
2.2 Semantic technology for context awareness
When modeling the real world according to the specific subject, ontology does not
only define clearly the concepts which came out after agreement from relevant
individuals and groups, but also be a logical group of terms which were expressed as
a formation that computers can understand and process (World Wide Web
Consortium, 2004).
Ontology defines the terms which are used in the domain and their relations, and it
is needed to share the knowledge information and the common semantic
understanding of its architecture. It is also used for efficient reusing of domain
knowledge. The characteristics of ontology is the plentiful expressions of relations
and causal relationship, connection and expansion of the real information and the
concept system, expansion of knowledge, providing reuse, mutual application and
connecting different information and knowledge base which is possible for logical
interface.
For ontology language, there are RDF (Resource Description Framework) and
OWL (Web Ontology Language). RDF is a standard language for the technology
exchanges of metadata which was presented by W3C. For mutual application among
various metadata, RDF provides common rules about semantics, architecture and
syntax (Study on the guidelines of web ontology development, National
Computerization Agency).
3. Research Methodology
Picture 2 is ontology constitutional factors. In this research, ontology is composed
with class, relations, Instances, axioms and cost value ontology. Class is concepts of
the domain or tasks, which are usually organized in taxonomies. Class in this
6
research is composed of user profile and spatial data. Relation is a type of interaction
between concepts of the domain. Instance is to represent specific elements of class.
Axiom is model sentences that are always true. This research uses a semantic
technique and applies for context-aware.
Picture 2) Ontology Constitutional factors
3.1. Context awareness based Ontology Inference System Architecture
The architecture of the total system which was suggested in this research is
comprised of ontology design part, user interface part and space database
part. Picture 3 is system architecture.
Picture 3) System Architecture
3.2 Space Preference
It is a part about important space preference among users’ context awareness.
According to the previous research which shows that users can use the services
from corresponding POI when the walking approach distance is within 400-500 m
7
and the walking hours is within maximum 6 minutes, users get the spatial database
they want within the maximum 500m.
3.3 Design User Profile
In order to use users’ context awareness, users’ basic information is inputted in the
system users use. In the user profile ontology, information about users’ age, sex,
possessing card, location, preference and preferred location is architected.
3.4 Design Relation
Through context awareness, time which is the most important variable for users’
context interpretation is an important attribute for users’ selection and the system.
It is because providing the most suitable schedule for users is the subject should be
preferentially considered when choosing the operating hours of service location or
movies from theater.
Picture 4) Time Table
As we can see in the picture 4, the user’s flexible hours is from 13 to 17 o’clock.
In case of movie 1, the starting time of the movie does not match with user’s flexible
time. In case of movie 3, it is not recommended to the user because the finishing
time of the movie does not match with user’s flexible time. However, in case of movie
2, the movie starts and finishes in the range of user’s flexible time. For this reason,
movie 2 is recommended to the users under the condition of time constraint.
8
3.5 Cost Value Ontology
In order to give weights on the selection criteria which are used when deciding POI
and service, cost value ontology was architected. In this study, distance rate and
discount rate were presented as selection criteria for POI and service. In order to
calculate weight, AHP Method was used.
In this study, discount rate and distance rate were selected as criteria factor of
selection. We decided the semantic scale and figure, and then organize them in a
matrix according to the decision criteria of selection. Through the pair wise
comparison between the criteria factor of selection and POI, we compare according
to the number of cases of every criteria factor of selection and POI.
Through the sum total of weight of selection criteria about for POI, the maximum
price is used for decision making, and this maximum price is applied to ontology.
3.6 Design Axiom
As knowledge which is expressed strictly, axiom is something we have to accept
without any proof of if it is right. In logic, public interest has an essential meaning to
show the knowledge which is the premise of interface.
When constructing ontology, axiom expresses strictly the definition of vocabulary
and concept. Axiom also has a function of interface about the questions of ontology
ability which is described using vocabulary and concept that are contained in
ontology.
3.7 Ontology
A mentioned above, through the users’ profile, space, constraint design, cost value
ontology, instance and axiom, the final ontology like below was architected. Picture 5
is ontology where is constructed in the last.
9
Picture 5) Ontology
4. System Application
In this chapter, we describe GIS interface system which unified the architected
ontology. Before materialization, we selected Myung-dong area which is developed
as various complex business districts because our subject region was downtown.
Picture 6 is a materialized system image. It is materialized to make it possible for
basic map view, map control, spatial query, geo-tagging and route finding.
Picture 6) GIS Interface System
10
In GIS interface system, there is a condition that users have to be in 400-500m of
walking approach distance and maximum 6 minutes of walking time in order to be able to
use services from the corresponding POI according to users’ space preference.
According to this condition, users can do space search for movie theaters which are
within 500m from the centered users, and search total POI which comes from the
corresponding space search.
We architect ontology using Protégé 4.1 – alpha for searched POI information and users’
context awareness.
As suitable service operators and services are interfaced to the users from the
architected ontology, at this time it is interfaced by using constraint conditions like
users’ basic information, preference, schedule and service hours.
Picture 7) Interface Result
In the picture 7, the interface result of the multi criteria decision is interfaced
through the ontology design system which is comprised of Protégé 4.1 – alpha of the
server. As decision making criteria is distance rate and discount rate, movie theater
IDs, 100003 and 100001 are interfaced according to each criteria. According to
users’ preference and current context, the recommended movie ‘Swon Brother’ was
interfaced. These interfaced information visualized the interface result through the
system users use.
11
Picture 8) Result Display
Picture 9) Route Finding
In the picture 8, the interface result is visualized through the map display and
information expression by GIS interface system. It shows the names of the
corresponding movie theaters and recommended movies, and their genre and movie
hours through the interfaced result as movie theater IDs. In the picture 9, it seeks the
routes to the recommended locations from the users’ current location.
Like the above materialization and experiment result, this system proved that it
could perform multi criteria decision making interface based on the users’ profile,
location information and space preference. The interface result could bring the
accurate result while performing location based personalized services to users.
5. Conclusion & Future direction
We suggested GIS interface system which recommends the suitable services to
users through the users’ context awareness and context interpretation.
In this research, we architected ontology which gives weight on users’ individual
characteristics, spatial geographic preference, service hours of POI and criteria
12
factors of decision making. After unifying interface functions and space analysis
functions, we showed the new location based service which brings the result by
interfacing suitable POI according to the users’ context awareness.
In order to verify the recommendation suitable for the individual users’ context, it
interfaced the movie schedule, current time, schedule according to the users’ flexible
hours, constraint conditions with the alliance cards of users’, recommended movie
theaters, recommended movies and movie hours from the public interest after
assuming the scenario of watching movie in the theater.
In this research, we architected ontology through the service information of users’
basic information, individual characteristics, preference, spatial geographic
preference information and POI. cost value ontology for decision making based on
users’ multi criteria was designed.
After unifying the visualized expression functions through the interface function for
multi criteria and spatial data management of GIS and spatial questions and the map
display of interface result using architected ontology and cost value ontology, we
presented the new LBS which brings out the result by interfacing the suitable POI
and service according to the users’ current context, basic information, preference
and space preference.
In the future researches, the study of tight coupling of ontology design part and GIS
system and Interface engine is needed to be processed, and more automatic and
intelligent functions are needed to be developed for the improvement of usefulness
for the suggested model.
Acknowledgements
This research was supported by a grant (06KLSGB01) from Cutting-edge Urban
Development – Korean Land Spatialization Research Project funded by Ministry of
Land, Transport and Maritime Affairs.
This work was researched by the supporting project to educate GIS experts.
13
References
YoungHoon Yu, JiHyeok Kim, Kwangcheol Shin, Geun Sik Jo, “Recommendation
system using location-based ontology on wireless internet: An example of collective
intelligence by using ‘mashup’ applications“, Expert Systems with Applications Vol
36(2009). (11675~11681)
Abolghasem Sadeghi Niaraki *, Kyehyun Kim, "Ontology based personalized route
planning system using a multi-criteria decision making approach", Expert System
with Applications Vol.36(2009). (2250~ 2259)
Shuain Yuan, Jun Shen, Jun Yan, "A Practical Geographic Ontology for Spatial
Web Services", 2008 IEEE International Conference on Services Computing
(579~580)
Antonio Coronato, Giuseppe De Pietro, Massimo Esposito, "A Sementic Location
Service for Pervasive Grids", OTM Workshops 2006, LNCS 4278 (1274~1294)
Bin Jiang, Xiaobai Yao, "Location-based services and GIS in perspective",
Computers Enviroment and Urban Systems 30 (2006). (715~725)
McGrath, R.E., Ranganathan, A., Campbell, R.H., Mickunas, M.D.: Use of
Ontologies in Pervasive Computing Environments. Technical report, University of
Illinois at Urbana-Champaign (2003)
Chen, H., Perich, F., Finin, T., Joshi, A.: SOUPA: Standard Ontology for Ubiquitous
and Pervasive Applications. In: International Conference on Mobile and Ubiquitous
Systems: Networking and Services (MobiQuitous), Boston, MA (2004)
Harry T. Uitermark, Peter J.M. van Oosterom, Nicolaas J.I. Mars, Martien Molenaar:
Ontology-based intergration of topographic data sets, International Journal of
Applied Earth Observation and Geoinformation 6 92005) 97-106
U. Visser, H. Stuckenschmidt, G. Schuster, T. Vogele: Ontoloies for geographic
information processing, Computers & Geosciences 28 (2002) 103-117
Femke Reitsma, John Laxton, Stuart Ballard, Werner Kuhn, Alia Abdelmoty:
Semantics, ontologies and eScience for the geosciences, Computers & Geosciences
35 (2009) 706-709
14
Alexander Zipf, Matthias Jost: Implementing adaptive mobile GI services based on
ontologies Examples from pedestrian navigation support, Computers Enviromment
and Urban Systems 30 (2006) 784-798
JaeGil Lee
2010 Dept. of GeoInformatic Engineering, In-ha University
Graduation (Bachelor of engineering)
2010~Present Dept. of GeoInformatic Engineering, In-ha
University Graduate school attending (Master`s course)
[Mailing Address] 253 Yonghyundong, Namgu, Incheon,
Republic of Korea 402-751 [Tel] +82-32-867-7605
[Fax] +82-32-863-1506 [Email] [email protected].
[Mobile] +82-10-2854-1370 [Research Interest] Geo-Ontology, Spatial Datamining,
Mobile GIS Service Model, Spatial Analysis.
YongJin Joo
2001 Dept. of GeoInformatic Engineering, In-ha University
Graduation (Bachelor of engineering)
2003 Dept. of GeoInformatic Engineering, In-ha University
Graduation (Master of Engineering))
2009. Dept. of GeoInformatic Engineering, In-ha University
Graduation (Ph D. GeoInformatic Engineering)
2003.1-2003.12 Software Developer, MobilCom
2004.1~2005.2 Researcher, Korea Transport Database, Korea Transport
Insititute(KOTI)
2009.11~Present Research Professor, Institute of Urban Sciences, University of
Seoul
[Mailing Address] Jeonnong-Dong, Dongdaemun-Gu, Seoul 130-743, Republic of
Korea [Tel] +82-2-2210-5253 [Fax] +82-2-2246-0186 [Email] [email protected]
15
[Mobile] +82-10-2701-7915 [Research Interest] GIS, Spatial DBMS, Telematics,
Embedded System, Urban Growth Simulation Model.
SooHong Park
1989. Seoul National University geography graduation
(bachelo of Geography)
1991. Seoul National University geography graduation
(master of Geography)
1996. Univ. of South Carolina at Columbia Graduation (Ph D.
geography)
1996.07~1997.08 Indiana University/ Research Associate
1997.09~1997.12 Seoul National University/researcher research study
1998.01~2000.02 Seoul Development Institute, Research committee
2000~Present Dept. of GeoInformatic Engineering, In-ha University, Associate
Professor
[Mailing Address] 253 Yonghyundong, Namgu, Incheon, Republic of Korea 402-751
[Tel] +82-32-860-7605 [Fax] +82-32-863-1506 [Email] [email protected].
[Mobile] +82-10-8970-9630 [Research Interest] u-GIS service model, spatial
databases, spatial data models.