2.1.3. Data Collection
Based on reading the 59 articles in the set, we have then collected information about each one. First of all, a set of metadata elements were collected, according to the following:
– Year of publication
– Number of authors
– Author affiliations
– Author nationality (according to the affiliation)
– Type of publication (i.e., journal article, book chapter, conference paper, or workshop paper)
In addition to these the topics of the articles were clas- sified along three dimensions; (i) theoretical or ap- plied research, (ii) main Semantic Web contribution, and (iii) type of DSS. Here we have chosen to classify something as theoretical research if the paper does not describe a system or other implementation, e.g., use case or empirical study, for the results. The category for applied research thereby represent those contribu- tions consisting of an implementation of some sort, ei- ther technical, such as a software system, or organiza- tional in terms of a case study or similar. In addition to these two categories we have singled out survey papers and position papers, which cannot normally be classi- fies into either of the two previous categories.
The main Semantic Web contribution has been clas- sified according to what Semantic Web technologies or approaches are at focus in the paper. The cate- gories were derived a-priori from the lists of confer- ence topics of the research track of the past four In- ternational Semantic Web Conferences. These topics have been quite stable, at least as far back as 2008, where only one new category has appeared (NLP - rep- resenting the increased hybridization of NLP and Se- mantic Web technologies) and one category has been replaced (“Applications of the Semantic Web” is re- placed with “Semantic Web Engineering”). The latter is most likely due to the presence of a specific “in use” track where applications are the main focus, hence, the research track now only targets the development rather than the applications themselves. We have cho- sen not to include the old “applications” category in our categorization, simply because we view the inter-section between DSS and Semantic Web technologies as one such application area, hence, all papers could be viewed as belonging to that category. Based on this, we end up with the following 6 categories, where the names and explanations have been slightly tailored to this study:
– Semantic Web data – representation languages, storage, search and querying of Semantic Web data, e.g., RDF data and Semantic Web Ser- vices, including approaches for using or produc- ing linked data, as well as quality assurance and provenance tracking of data.
– Ontologies and semantics – representation lan- guages and patterns, engineering, management, retrieval and usage of Semantic Web ontologies and rules, including reasoning services and rule execution engines.
– Semantic Web engineering and development - building of Semantic Web applications, methods, tools and evaluations of applications.
– Natural Language Processing (NLP) – machine learning and information extraction for the Se- mantic Web, Semantic Web population from text or from exploiting tags and keywords, or using semantic technologies to perform NLP.
– Social Semantic Web – social networks and pro- cesses, collaboration and cooperation, context awareness and user modelling, trust, privacy, and security.
– User interfaces – interaction with and creation of Semantic Web data and models, information pre- sentation, visualization and integration, personal- ization.
To exemplify how these categories have been assigned, assume a paper presenting an approach to use ontolo- gies for information integration and subsequently pre- senting that information using a novel visualization method exploiting the underlying semantics of the in- formation. Such a paper would be classified as belong- ing to both the second and final category of the list above, while an application using ontologies for infor- mation integration, but using standard user interface components not tailored to the Semantic Web, would be classified as belonging only to the second category of the list. Similarly a paper discussing privacy and security has only been assigned the social Semantic Web category if Semantic Web-related technologies in some way contribute to those issues or their solution. Hence, we have only classified the contribution of the Semantic Web technologies used in the paper into these categories, not the overall approach of the paper, and each paper may have multiple classifications.
Finally, we have studied the type of DSS addressed in the paper. For this we have used the categories of DSS listed in [11], as presented in the bullet list of Section 1.1. We have slightly adapted one of the cate- gory definitions, compared to [11]; a personal DSS is not restricted to a DSS that is tailored for, or used by, only a small number of users, instead we define a per- sonal DSS as being targeted tow
2.1.3 2.1.3. Data Collection
Based on reading the 59 articles in the set, we have then collected information about each one. First of all, a set of metadata elements were collected, according to the following:
– Year of publication
– Number of authors
– Author affiliations
– Author nationality (according to the affiliation)
– Type of publication (i.e., journal article, book chapter, conference paper, or workshop paper)
In addition to these the topics of the articles were clas- sified along three dimensions; (i) theoretical or ap- plied research, (ii) main Semantic Web contribution, and (iii) type of DSS. Here we have chosen to classify something as theoretical research if the paper does not describe a system or other implementation, e.g., use case or empirical study, for the results. The category for applied research thereby represent those contribu- tions consisting of an implementation of some sort, ei- ther technical, such as a software system, or organiza- tional in terms of a case study or similar. In addition to these two categories we have singled out survey papers and position papers, which cannot normally be classi- fies into either of the two previous categories.
The main Semantic Web contribution has been clas- sified according to what Semantic Web technologies or approaches are at focus in the paper. The cate- gories were derived a-priori from the lists of confer- ence topics of the research track of the past four In- ternational Semantic Web Conferences. These topics have been quite stable, at least as far back as 2008, where only one new category has appeared (NLP - rep- resenting the increased hybridization of NLP and Se- mantic Web technologies) and one category has been replaced (“Applications of the Semantic Web” is re- placed with “Semantic Web Engineering”). The latter is most likely due to the presence of a specific “in use” track where applications are the main focus, hence, the research track now only targets the development rather than the applications themselves. We have cho- sen not to include the old “applications” category in our categorization, simply because we view the inter-section between DSS and Semantic Web technologies as one such application area, hence, all papers could be viewed as belonging to that category. Based on this, we end up with the following 6 categories, where the names and explanations have been slightly tailored to this study:
– Semantic Web data – representation languages, storage, search and querying of Semantic Web data, e.g., RDF data and Semantic Web Ser- vices, including approaches for using or produc- ing linked data, as well as quality assurance and provenance tracking of data.
– Ontologies and semantics – representation lan- guages and patterns, engineering, management, retrieval and usage of Semantic Web ontologies and rules, including reasoning services and rule execution engines.
– Semantic Web engineering and development - building of Semantic Web applications, methods, tools and evaluations of applications.
– Natural Language Processing (NLP) – machine learning and information extraction for the Se- mantic Web, Semantic Web population from text or from exploiting tags and keywords, or using semantic technologies to perform NLP.
– Social Semantic Web – social networks and pro- cesses, collaboration and cooperation, context awareness and user modelling, trust, privacy, and security.
– User interfaces – interaction with and creation of Semantic Web data and models, information pre- sentation, visualization and integration, personal- ization.
To exemplify how these categories have been assigned, assume a paper presenting an approach to use ontolo- gies for information integration and subsequently pre- senting that information using a novel visualization method exploiting the underlying semantics of the in- formation. Such a paper would be classified as belong- ing to both the second and final category of the list above, while an application using ontologies for infor- mation integration, but using standard user interface components not tailored to the Semantic Web, would be classified as belonging only to the second category of the list. Similarly a paper discussing privacy and security has only been assigned the social Semantic Web category if Semantic Web-related technologies in some way contribute to those issues or their solution. Hence, we have only classified the contribution of the Semantic Web technologies used in the paper into these categories, not the overall approach of the paper, and each paper may have multiple classifications.
Finally, we have studied the type of DSS addressed in the paper. For this we have used the categories of DSS listed in [11], as presented in the bullet list of Section 1.1. We have slightly adapted one of the cate- gory definitions, compared to [11]; a personal DSS is not restricted to a DSS that is tailored for, or used by, only a small number of users, instead we define a per- sonal DSS as being targeted tow
การแปล กรุณารอสักครู่..
