@InProceedings{BertonLope:2016:NeCoAp,
               author = "Berton, Lilian and Lopes, Alneu de Andrade",
          affiliation = "{Universidade do Estado de Santa Catarina} and {Universidade de 
                         S{\~a}o Paulo}",
                title = "Network Construction and Applications for Semi-Supervised 
                         Learning",
            booktitle = "Proceedings...",
                 year = "2016",
               editor = "Aliaga, Daniel G. and Davis, Larry S. and Farias, Ricardo C. and 
                         Fernandes, Leandro A. F. and Gibson, Stuart J. and Giraldi, Gilson 
                         A. and Gois, Jo{\~a}o Paulo and Maciel, Anderson and Menotti, 
                         David and Miranda, Paulo A. V. and Musse, Soraia and Namikawa, 
                         Laercio and Pamplona, Mauricio and Papa, Jo{\~a}o Paulo and 
                         Santos, Jefersson dos and Schwartz, William Robson and Thomaz, 
                         Carlos E.",
         organization = "Conference on Graphics, Patterns and Images, 29. (SIBGRAPI)",
            publisher = "Sociedade Brasileira de Computa{\c{c}}{\~a}o",
              address = "Porto Alegre",
             keywords = "network construction, graph-based methods, semi-supervised 
                         learning, complex networks.",
             abstract = "The influence of network construction on graph-based 
                         semi-supervised learning (SSL) and their related applications have 
                         only received limited study despite its critical impact on 
                         accuracy. We introduce four variants for networkconstruction for 
                         SSL that adopt different network topology: 1) S-kNN (Sequential 
                         k-Nearest Neighbors) that generates regular networks; 2) GBILI 
                         (Graph Based on the informativeness of Labeled Instances) and 3) 
                         RGCLI (Robust Graph that Considers Labeled Instances), which 
                         exploit the labels available generating scale-free networks; 4) 
                         GBLP (Graph Based on Link Prediction), which are based on link 
                         prediction measures and creates smallworld networks. Comprehensive 
                         experimental results using several benchmark datasets show that it 
                         can achieve or outperform existing state-of-the-art results. 
                         Furthermore, it is confirmed to be more effective in running 
                         time.",
  conference-location = "S{\~a}o Jos{\'e} dos Campos, SP, Brazil",
      conference-year = "4-7 Oct. 2016",
             language = "en",
                  ibi = "8JMKD3MGPAW/3M9H4AE",
                  url = "http://urlib.net/ibi/8JMKD3MGPAW/3M9H4AE",
           targetfile = "WTD-SIBGRAPI2016-LilianBerton-2.pdf",
        urlaccessdate = "2024, Apr. 28"
}