Network aware team formation

From Master Projects
Jump to: navigation, search

has title::Network aware team formation
status: finished
Master: project within::Computational Intelligence and Selforganisation
Student name: student name::Bart Jan van Engelen
Start start date:=2011/09/01
End end date:=2012/08/22
Supervisor: Lora Aroyo
Second supervisor: Rory Sie
Second reader: has second reader::Rory Sie
Company: has company::OU and VU
Thesis: has thesis::Media:Thesis_Network_aware_team_formation_BJvEngelen_1329685.pdf
Poster: has poster::Media:Thesis Proposal L Aroyo bengelen 1329685.pdf

Signature supervisor



Utilizing publicly available social collaboration network data of one university we propose to build a recommender system that assists researchers in selecting their research partners. We hypothesize that the influence an author has on the scientific collaboration network as well as the similarities of interest between two authors are predictors for the influence of the scientific output created by these authors. Furthermore we expect that making the network in which researchers co-operate more transparent by means of the recommender is of it self of value to the researchers. Although we did not find definitive proof that the recommender selects valuable co-operations above less valuable co-operations in our dataset we did find evidence that at least the influence of an author on the co- operation network is a good predictor for the influence of the scientific output of this author. What is more we find that users are moderately positive towards the recommendations of the system and find the system easy to use. Lastly we see that the network of our university is comparable to that of other universities, giving some external validity to these results.