An Algorithm to Generate Data Linkages of Linked Sensor Data
Lin-na Cuan
1
, Yi-min Shi
2
, Guan-yu Li
1
, Xue-hua Wu
1
, Ning Liu
2
Department of Information Science & Technology
Dalian Maritime University
Dalian, China
1
e-mail: {cuanlinna120908, rabitlee, wuxuehua00111} @163.com
2
e-mail: {shiyimin1966, liun571}@126.com
Abstract—To form Linked Sensor Data, we can generate
linkages among the sensor network data described by semantic
information and relevant resources in a Linked Open Data
cloud, which can make efficient use of sensor network data.
Based on the analysis of existing interlinking methods of
Linked Data and Linked Sensor Data publishing systems,
according to the property characteristics of sensor network
data as well, we propose a similarity comparison algorithm of
graph based on heuristic property, to generate linkages among
sensor network data and related resources in Linked Open
Data cloud effectively.
Keywords-Linked Data; Linked Sensor Data; linkages;
heuristic property; RDF graph
I. INTRODUCTION
The goal of Linked Data [1] is to enable people and
organizations to share the structured data on the Web as
easily as they can share documents today. The most
important value of it is “links” which supports arbitrary
association of structured data, and navigating the entire
network. People can use Linked Data browsers to travel from
a data source to another through RDF links, to get more and
more comprehensive information. Generating data linkages
among relevant resources is the premise of Linked Data
building, publishing and extending. At present, the existing
automatic interlinking approaches of Linked Data include:
mapping based on entity’s text [2], graph’s similarity [3] and
rules[4]. Mapping based on entity’s text is the basic method
and mapping based on graph’s similarity is an extension of
single triple comparison. These two approaches are general
and common methods, but the relationships they can create
are very limited. The approach of interlinking based on rules
can create richer and more complex relationship types, but it
depends on specific data models and related rules.
Because of the sensor network data stored in different
formats and lacking semantic descriptions, it cannot be
efficiently used by applications, to solve this problem,
Linked Sensor Data [5] is proposed. The key problem of
Linked Sensor Data generating process which builds
linkages among sensor network data described by semantic
information and related data in Linked Open Data (LOD)
cloud, is whether we can use the automatic interlinking
method of Linked Data to generate linkages among sensor
data and its related resources. Gotten from sensing device to
detect changes of environment, sensor network data is a
relational database and has observed property characteristic.
To generate linkages among sensor network data and its
related resources, we should not only use automatic
interlinking approach of Linked Data, but also compare
similarity of their property concepts. Learning from the ideas
of mapping based on graph’s similarity method of Linked
Data and setting the observed property of sensor network
data as an inspiration, we propose a similarity comparison
algorithm of graph based on heuristic property (SCAG-HP),
to generate linkages among sensor network data and related
resources in LOD cloud.
II. B
ACKGROUND
A. Linked Data
The term Linked Data was coined by Tim Berners-Lee in
2006 [1]. Linked Data is the best data linking method for
exposing, sharing, and connecting pieces of data,
information, and knowledge on the Semantic web using
URIs and RDF. It also emphasizes interlinking of the data
and contextual information useful for human-computer
understanding.
The basic tenets of Linked Data can be described as
follows [1]:
• Using URIs as names for things,
• Providing HTTP access to those URIs,
• Providing useful information for URIs using the
standards (RDF, SPARQL),
• Including links to other URIs.
Linked Open Data cloud(LOD)[1,6]
is to publish the
open data source on the Web in the form of RDF, such as
Wikipedia, GeoNames, WordNet etc. Meanwhile it generates
the RDF links among the data sources for browsers, search
engines and more advanced applications of Linked Data.
LOD cloud has already collected 203 datasets, 25 billion
RDF triples and 395 million RDF links. These data sets
include the information in the geography, life sciences,
medicine, publishing, media, social networks, and others.
Therefore, Linked Data has been widely used in many fields
such as politics, business and biology. For example, BBC
(British Broadcasting Corporation) uses the Linked Data
technology to achieve the permanent link and openness for
data of each program [7]. Similarly, Data.gov provides a
service based on Linked Data, accomplishing navigation and
acquisition of geographical entities within the territory of the
United Kingdom [8].
2015 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery
978-1-4673-9200-6/15 $31.00 © 2015 IEEE
DOI 10.1109/CyberC.2015.105
380