Finding and Exploring Commonalities between Researchers Using the ResXplorer
1 like940 views
The document describes a tool called ResXplorer that helps researchers explore commonalities between their work and other related publications, people, and events. It does this by analyzing a researcher's social media profile and publications against a linked open data knowledge base. ResXplorer provides an interactive radial interface where related entities are displayed based on their affinity ranking. An evaluation showed ResXplorer can efficiently find commonalities between researchers with high precision, and users perceived it as useful for exploring relationships in research data.
Finding and Exploring Commonalities between Researchers Using the ResXplorer
- 1. Finding and Exploring Commonali0es between Researchers Using the ResXplorer HCII2014, Crete, Greece Selver SoEic, Laurens De Vocht, Mar0n Ebner, Erik Mannens, Rik Van de Walle
- 2. Mo0va0on and Challenges • Vast of scien0fic publica0ons available online – proprietary interfaces with poor context overview – costs for full access required • Researchers – use Social Media to communicate about their work – have interests on events, persons and resources from research related to them • Digital publica0on archives available as Linked Data • Lack of analy0c services and adequate visualiza0ons
- 3. Goals Make the most of the wealth of resources for research: through rela0ng and aligning scholar profiles with the online available resources, publica0ons, and other scholar's profiles. Provide a interface for visual explora0on and search based upon commonali0es derived from this alignment
- 6. Finding Commonali0es • Real-‐0me keyword disambigua0on of entry results • Finding rela0ons between the entered keywords through the EiCE • Addi0onally enhancement – social media profile as search entry point – Expansion of results over radial interface
- 7. ResXplorer • “ResXplorer” is aggregated interface for search and explora0on of the underlying Linked Data Knowledge Base • Radial Interface with clickable en00es differing through color, size and shape • Resides on EiCE Engine ( our prior work on path finding between resources in seman0c en0ty graphs) • Interface adapts the explora0on state and focused en0ty is always central. The distance and posi0on of other en00es is defined by affinity ranking algorithm. http://www.resxplorer.org
- 8. Evalua0on
- 10. Measure Links discovered along traversing path of algorithm which lead to scientific resources (publications, persons and events) relevant for one of the both authors represent true positives.
- 11. 0 5 10 15 20 25 30 35 40 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90 1,00 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Precision # Commonali0es Precision vs. #Commonali0es
- 12. 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 # Commonali0es # TP Commonali0es # FP Commonali0es #Commonali0es vs. #TP vs. #FP
- 13. 0 1 2 3 4 5 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90 1,00 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Precision Path length Precision vs. Pathlength
- 14. 0 5 10 15 20 25 30 35 40 0 1 2 3 4 5 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Path length # Commonali0es Pathlength vs. #Commonali0es
- 15. Front-‐End Evalua0on • TAM Model (Davis 1985) – perceived usefulness • Short Survey with 31 researcher – all from computer science domain • Likert Scala: strongly agree (5) – strongly disagree (1)
- 17. Conclusions • ResXplorer allows researchers to interac0vely explore rela0ons between the resources and en00es like events, places, publica0ons or persons related to their work and discover commonali0es between them • Finding commonali0es tends to reach high precision which does not depend from the length of search path, and the count of found commonality links • ResXplorer interface is according to “percieved usefullness” survey made with target users meant primarily to explore rela0on between research en00es
- 18. Related Work • SoEic, Selver; Ebner, Mar0n; De Vocht, Laurens, Mannens, Erik, Van de Walle, Rik,,A Framework Concept for Profiling Researchers on Twiker using the Web of Data,9th Interna0onal Conference on Web Informa0on Systems and Technologies, pp. 447 -‐ 452, Aachen, Germany, 2013 • De Vocht, Laurens, SoEic, Selver, Ebner, Mar0n, Mühlburger, Herbert, ,Seman0cally driven social data aggrega0on interfaces for Research 2.0,Proceedings of the 11th Interna0onal Conference on Knowledge Management and Knowledge Technologies,,Ar0cle 43,1-‐9,2011 • De Vocht, L., Coppens, S., Verborgh, R., Vander Sande, M., Mannens, E., & Van de Walle, R. (2013). Discovering meaningful connec0ons between resources in the web of data. 6th Workshop on Linked Data on the Web, Proceedings. Presented at the 6th Workshop on Linked Data on the Web. • De Vocht, Laurens, SoEic, Selver, Mannens, Erik, Van de Walle, Rik, Ebner, Mar0n, A Search Interface for Researchers to Explore Affini0es in a Linked Data Knowledge Base,The 12th Interna0onal Seman0c Web Conference (ISWC 2013),Posters and Demo Session,2013,CEUR-‐WS • De Vocht L., SoEic S., Mannens E., Van de Walle R., Ebner M., ResXplorer: Interac0ve Search for Rela0onships in Research Repositories, Seman0c Web Challenge, part of the 12th Interna0onal Seman0c Web Conference, Sydney,Australia 2013 • De Vocht L., SoEic S., Mannens E., Van der Walle R., Ebner M., Aligning Web Collabora0on Tools with Research Data for Scholars, Proceedings of the Bigscholar Workshop on World Wide Web conference 2014 (WWW2014), pp.1203-‐1208, Seoul, Korea, 2014. • De Vocht L., SoEic S., Mannens E., Van der Walle R., Ebner M., Finding Useful Resources for Researchers within the Web of Data, Journal of Web Seman0cs, Special Issue on Seman0c Search, 2014.