Document Actions
Bibex: Bibliographic Exploration
Basic Description
BIBEX is based on the graph database query engine DEX and the information extracted from the DBLP. BIBEX allows for complex bibliographic search and shows the results of its queries as a combination of graphs and text for two types of scenarios: scientists who want to do complex bibliographic search, and PC members and journal editors who want to have an aid to search best suited reviewers that do not have conflicts of interest with the authors of papers submitted for review. The tool allows to explore the relation between authors, keywords and papers in such a way that a user may perform a complex bibliographic search to investigate the who is who in specific areas of research.
Introduction
The need for powerful bibliographic databases and search tools has motivated the existence of DBLP, Google Scholar or Citeseer. The importance of those devices has been evident and many scientists are using them regularly to do bibliographic keyword search. However, the graph nature of the data stored in those databases has not been exploited at its most. Such nature, would allow to place complex queries leading to answers with an added value to the scientific community.
BIBEX stands for Bibliographic Exploration Tool and has the objective to answer pre-programmed queries over the data sets at hand. An answer to a query with BIBEX is basically a graph of relations plus a set of descriptive and/or statistical information that can be obtained from the answers to the query or the graph database itself. In order to create a full graph repository, BIBEX links the records in the databases on the basis of the relationships among them. The relationships between records are explicit within each database, or artificially created among different databases using the names of the authors, the keywords of the papers or the names of the papers. BIBEX allows to perform queries under two different flavors: (i) that of the researcher who intends to search for papers and their references, and authors and their relationships in specific areas of research, and (ii) that of the journal editor or PC member who search for good reviewers for a paper, i.e. non interest-conflicting researchers. There are different graph database query engine proposals in the literature that could be used for our purpose. However, DEX has some interesting characteristics that make it suitable for the bibliographic context: it allows to deal with multiple data sources in different formats, e.g. relational, XML, CSV; it is oriented to deal with large data sets and it provides fast response times.
The graph shows author names in the nodes with links between their collaborators and tags in the links with the number of papers written by them. The query allows to include topic or paper keywords as parameters. At this moment we use the paper title keywords for this purpose but the search could be enriched with real keywords present in other sources. The use of keywords allows to restrict the search to specific topic areas for the author or authors analyzed. A query generates a view with all the subgraphs that respond to it, or one view for each of the subgraphs generated.
Parameters:
< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: On mouse left button click on a node or edge of the answer graph, BIBEX generates information in the right hand side panels:
Use of the query: This query may be used to investigate authors and their relation to topic areas, the papers they have written and who they have collaborated with. This is useful for researchers who are new in an area like PhD students, or those who want to know new papers in a topic.
The result for this query shows a graph for each set of keywords with a minimum support, showing the authors related to those keywords.
Parameters:
< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: The mouse options for this query are similar to those in query 1, and follow the same logics. For these reasons we avoid explaining them.
Use of the query: This query may be used to search for papers in areas of research when the user does not have a certain knowledge of the keywords in the area. Also, when one wants to know about the papers published close to a certain area.
< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: The mouse options for this query are similar to those in query 1, and follow the same logics. For these reasons we avoid explaining them.
Use of the query: This query, may be used by editors and PC members to find reviewers for papers. It may be also helpful for researchers looking for groups that work on similar topics in a competitive way.
Gómez-Villamor, S., Soldevila-Miranda, G., Giménez-Vañó, A., Martínez-Bazan, N., Muntés-Mulero, V., and Larriba-Pey, J. 2008. In Proceedings of the 11th international Conference on Extending Database Technology: Advances in Database Technology (Nantes, France, March 25 - 29, 2008). EDBT '08, vol. 261. ACM, New York, NY, 735-739.
Bibex Interface
The interface of BIBEX is very simple, including six main panels. The top left panel lists all the available queries that may be launched. Nowadays, three queries are available. Once a query is selected, the same panel allows to set the parameters for the specified query. The bottom left panel shows a tree with the history of executed queries. The central panel shows the main answer to a query in the form of a graph. The right hand side panels show specific information for a selected component (node or edge) from the graph of the central panel. Specifically, the top right panel shows the information integrated from different data sources or the selected component, the middle right panel shows inferred papers from the query, and the bottom right panel shows statistics gathered during the execution of the query. The system allows for simultaneous live queries. This is important since it may be necessary to have multiple simultaneous answers for a specific bibliographic investigation. The tool also allows to obtain printed reports that reflect a condensed summary of a query. With this information, the search process may be easier to be carried. Finally, the system allows to navigate and obtain the papers from the bibliographic sources when a url exists and the user is granted access to those sources. It is also possible to access the web pages of the authors if the information is included in the databases.Queries
We divide the queries into two different sets, those for the plain researcher, and those for the editor/PC member.Researcher queries
Query 1. Partners/keywords.
This query generates a graph with all the authors that have collaborated with one or more authors specified in the query.The graph shows author names in the nodes with links between their collaborators and tags in the links with the number of papers written by them. The query allows to include topic or paper keywords as parameters. At this moment we use the paper title keywords for this purpose but the search could be enriched with real keywords present in other sources. The use of keywords allows to restrict the search to specific topic areas for the author or authors analyzed. A query generates a view with all the subgraphs that respond to it, or one view for each of the subgraphs generated.
Parameters:
< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: On mouse left button click on a node or edge of the answer graph, BIBEX generates information in the right hand side panels:
- The stored data for the node (top panel).
- The papers related to the edge or node (middle panel).
- The statistics gathered for the query (bottom panel).
Use of the query: This query may be used to investigate authors and their relation to topic areas, the papers they have written and who they have collaborated with. This is useful for researchers who are new in an area like PhD students, or those who want to know new papers in a topic.
Query 2. Keyword recommender.
Based on a set of authors and keywords, the query searches for the keywords that have not been specified in the query, but are present in the papers involved in the query. Among those keywords, it gathers the groups of keywords that are above a certain support threshold. Finally, it searches for the papers that contain the groups of supported keywords and shows the relation among the authors who have written them.The result for this query shows a graph for each set of keywords with a minimum support, showing the authors related to those keywords.
Parameters:
< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: The mouse options for this query are similar to those in query 1, and follow the same logics. For these reasons we avoid explaining them.
Use of the query: This query may be used to search for papers in areas of research when the user does not have a certain knowledge of the keywords in the area. Also, when one wants to know about the papers published close to a certain area.
Editor/PC Member queries
Query 3. Reviewers for a paper.
Based on a set of keywords and author names, this query recommends a set of reviewers that do not have any conflict of interest with the authors. Information related to authorship is used to discard researchers that have collaborated with the authors stated in the query. The query returns a graph with researchers that could review a paper related to the set of keywords introduced for the query, and their relations with other researchers.
Parameters:< author-name1, author-name2, ..., author-nameN >
< keyword1, keyword2, ..., keywordN >
Options: The mouse options for this query are similar to those in query 1, and follow the same logics. For these reasons we avoid explaining them.
Use of the query: This query, may be used by editors and PC members to find reviewers for papers. It may be also helpful for researchers looking for groups that work on similar topics in a competitive way.
Usage Demo
Related Bibliography
Gómez-Villamor, S., Soldevila-Miranda, G., Giménez-Vañó, A., Martínez-Bazan, N., Muntés-Mulero, V., and Larriba-Pey, J. 2008. In Proceedings of the 11th international Conference on Extending Database Technology: Advances in Database Technology (Nantes, France, March 25 - 29, 2008). EDBT '08, vol. 261. ACM, New York, NY, 735-739.
