New Trends in Semantic Web-Based Applications
J.UCS Special Issue
Miguel Ángel Rodríguez-García
(Universidad Rey Juan Carlos, Madrid, Spain
miguel.rodriguez@urjc.es)
Rafael Valencia-García
(Universidad de Murcia, Spain
valencia@um.es)
Giner Alor-Hernández
(Tecnológico Nacional de México/I. T. Orizaba, Orizaba, Veracruz, México
galor@itorizaba.edu.mx)
The Semantic Web brings a new way of representing information, thereby
making it more understandable for computers. Ontologies are the
building blocks of the semantic web framework and have thus become its
cornerstone, as they allow sharing and reusing - in a standardized way
- heterogeneous data between applications of different nature. In this
sense, ontology capabilities have paved the way for new, innovative
trends in web-based applications across multiple domains, including
medicine, e-learning, and ontology and software engineering.
This Special Issue explores how semantic web application trends are
being developed as of the popularity of semantic web technologies. To
prepare this compilation of works, the editors promoted an open call
for papers through multiple international email lists and received a
vast number of prospects. Then, every paper was thoroughly
peer-reviewed by subject matter experts. The peer-review process
resulted in the selection of eight high-quality manuscripts compiled
and published in this Special Issue.
In the first paper, titled Astmapp: A Platform for Asthma
Self-Management, authors Harry Luna-Aveiga, José Medina-Moreira, Oscar
Apolinario-Arzube, Mario Andrés Paredes-Valverde, Katty Lagos-Ortiz
and Rafael Valencia-García introduce Astmapp, a platform for
supporting the asthma self-management process. The platform relies on
semantic technologies to define an ontology that models asthma
symptoms, patient behaviour and mood, asthma triggers, and
concepts. Similarly, mobile technologies are employed to record
patient activity, whereas recommender systems provide patients with a
direct access line to an asthma-related knowledge repository.
In the second manuscript, named EduRP: An Educational Resources
Platform based on Opinion Mining and Semantic Web, Maritza Bustos
López, Giner Alor-Hernández, José Luis Sánchez-Cervantes, María del
Pilar Salas-Zárate, and Mario Andrés Paredes-Valverde propose an
e-learning platform that is primarily based on semantic profiling and
sentiment analysis. Specifically, semantic profiling is used to build
semantic user profiles, whereas sentiment analysis helps assess the
quality of every educational resource.
To validate the platform's performance, the authors conducted
experiments by using a labelled corpus of educational resources. The
harvested results confirmed the platform's high accuracy when
detecting the polarity of Spanish educational resource reviews.
In the third work, Enabling System Artefact Exchange and Selection
through a Linked Data Layer, researchers Jose María Alvarez-Rodríguez,
Roy Mendieta, Jose Luis de la Vara, Anabel Fraga, and Juan Llorens
tackle the problem of artefact reusability across the system
development lifecycle stages. Generally, these artefacts are created
along the process in a private way and in a different
format. Therefore, the authors propose a real multi-format system
based on Open Services for Lifecycle Collaboration (OSLC), which
enables developers to share and exchange any artefact defined under
the Linked Data initiative. The proposed solution is validated against
other platforms, and the authors obtain significant improvements.
In the fourth manuscript, named Enhancing Spatial Keyword Preference
Query with Linked Open Data, João Paulo Dias de Almeida, Frederico
Aráujo Durão, and Arthur Fortes da Costa present an algorithm that
increases the accuracy of spatial keyword preference querying. The
algorithm utilizes Linked Open Data to represent the keywords and
features used to describe objects located in the repository. Then,
given a set of keywords, the authors use SPARQL to explore their
neighbourhood and extract more details, which are used to enrich the
search. To assess the algorithm, the authors conduct two different
experiments, thus comparing the algorithm's accuracy against that
of the traditional spatial keyword preference query.
In the fifth paper, Shadi Abudalfa and Moataz Ahmed discuss their
work: Open Domain Targeted Sentiment Classification Using
Semi-Supervised Dynamic Generation of Feature Attributes. Namely, the
researchers propose a new semi-supervised learning technique for
opinion mining that requires small amounts of labelled data. The
method utilizes hidden Markov Support Vector Machine (HM-SVM) to
improve the accuracy of the sentiment classification. Finally, to
assess the novel technique, Shadi and Moataz conducted multiple
experiments to compare the efficiency of their technique with that of
previous related works.
In the sixth contribution, titled Mining of Educational Opinions with
Deep Learning, researchers Raúl Oramas Bustillos, Ramón Zatarain
Cabada, and María Lucía Barrón Estrada describe an opinion mining
module that relies on deep learning techniques to accomplish sentiment
analysis on student opinions about exercises in Java. As a part of
their conclusions, the authors conduct multiple experiments to assess
the effectiveness of the deep learning module.
In the seventh paper, Laura Po and Davide Malvezzi introduce their
work: Community Detection Applied on Big Linked Data. Namely, the
authors describe high-level visualizations on Big Open Linked Data
(known as H-BOLD), a new tool that enables users to explore Big Linked
Data. As its main characteristic, H-BOLD utilises schema summary and
community detection techniques to represent and visualize RDF
datasets. Similarly, the authors assess H-BOLD's performance by
comparing four different community detection algorithms aimed at
finding potentially collapsible nodes, taking into account the density
of their connections.
Finally, in Human Language Technologies: Key Issues for Representing
Knowledge from Textual Information, Yoan Gutiérrez, Elena Lloret, and
José M. Gómez describe an ontology schema that has been defined to
semantically express text content. Furthermore, the authors define the
algorithm called Semantic Package population to populate the
pre-defined ontology schema. To conclude the work, the researchers
evaluate the ontology schema by conducting several tests that analyse
concrete features about its structure and organization.
The Special Issue editors would like to thank the authors for
submitting their papers, as well as the reviewers for their invaluable
comments and suggestions. Special thanks also to Professor Christian
Gütl (Managing Editor) and Ms Dana Kaiser (Assistant Editor), from
the Journal of Universal Computer Science (J.UCS), for their support
and the opportunity to edit this work.
Acknowledgements
This special issue has been supported by the Spanish National Research
Agency (AEI) and the European Regional Development Fund (FEDER / ERDF)
through project KBS4FIA (TIN2016-76323-R) and the Community of Madrid
through the Talent Attraction Program
("2017-T2/TIC-5664"). Additionally, the issue is supported by
Tecnológico Nacional de México (TecNM) and sponsored by Mexico's
National Council of Science and Technology (CONACYT) and the
Secretariat of Public Education (SEP) through PRODEP (Programa para el
Desarrollo Profesional Docente, for its acronym in Spanish).
Miguel Ángel Rodríguez-García
Rafael Valencia-García
Giner Alor-Hernández
|