2. Literature Survey
In order to characterize the research that has been conducted on Semantic Web technologies for DSS so far, we have performed a structured literature review, which is described in this section.
2.1. Structured Method
To generate a manageable but comprehensive set of articles for our survey, we have chosen to target re- search publications where the authors explicitly claim to work within this intersection. Later, in Sect. 4 we will discuss the research areas more broadly, and draw on our own knowledge and experience from the Se-mantic Web field, but in this part of the paper we take a more structured approach. Below we describe first the literature collection method, next, the data collection performed based on the collected literature, and subse- quently in the following section we present the results of the structured literature survey.
2.1.1. Literature Collection
In order to make an unbiased selection, and gener- ate a reasonable coverage of all literature that explic- itly claims to treat the intersection of DSS and Se- mantic Web research, we first selected a number of keywords representing each of these fields. The DSS field was here represented by the two keywords deci- sion support and business intelligence, the second one selected because it is sometimes used as a synonym for decision support in the more business-focused lit- erature. The Semantic Web field was then represented by the keywords Semantic Web, semantic technologies, linked data, and any combination of ontology with ei- ther RDF or OWL. Apart from the first two, which are quite obvious, we wanted to capture articles that used some of the more prominent technologies of the field, but without actually mentioning their origin in the Se- mantic Web. For each source (see further below) arti- cles were retrieved that contained any combination of a DSS keyword and a Semantic Web one, i.e., resulting in up to 10 distinct searches being made within each source.
The sources are of three main types; (i) general online indexing service, (ii) publisher database, and (iii) individual journal or publication series. As a representative of the first category, we used Google Scholar, which indexes a multitude of online publi- cation databases from various publishers. Due to the huge number of articles indexed, we searched for the presence of the 10 combinations of keywords only in article titles and abstracts. Representing the second category, we used the SpringerLink online database, since it covers many of the publications within the Se- mantic Web field, e.g., proceedings of the most promi- nent conferences such as ISWC and ESWC. Also in this case, the total number of indexed articles is huge, hence, we restricted the search to articles with any of the 10 keyword combinations in the title or abstract.
Finally, representing the third category, we selected a number of journals from both fields. Here the Seman- tic Web field is being represented through the Journal of Web Semantics (Elsevier) and the Semantic Web Journal (IOS Press). Previous reviews of the DSS field, e.g., [11], have listed the most prominent journals in DSS research, whereas based on these results we chose to focus on the Decision Support Systems journal (El- sevier), Decision Sciences (Wiley), the International Journal of Information Technology and Decision Mak- ing (World Scientific), Information and Management (Elsevier), International Journal of Spatial Data Infras- tructures Research (online journal by the Joint Re- search Centre of the European Commission), and MIS Quarterly (University of Minnesota). The respective online search facilities of the listed journals were used for retrieving articles, by means of the keywords de- scribed above. Depending on the facilities provided by the respective sites, if present we included a full-text search of the article content in addition to searching title and abstract. For searching the specific journals we made an additional assumption; all articles of the two Semantic Web journals were assumed to be about the Semantic Web, hence only the two DSS-related keywords were used for retrieval, and the opposite for the clearly DSS-related journals (Decision Sup- port Systems and Decision Sciences), where only the 5 keyword combinations representing Semantic Web concepts were used, while all the remaining IS jour- nals were treated similarly to the general databases. All articles that were either directly available online, or retrievable through library order were collected. A few articles of Google Scholar turned out to constitute “broken links”, and hence were not retrieved.