“School of Cognitive Sciences”

Back to Papers Home
Back to Papers of School of Cognitive Sciences

Paper   IPM / Cognitive Sciences / 13148
School of Cognitive Sciences
  Title:   Combining classifiers using nearest decision prototypes
  Author(s): 
1.  S.R. Kheradpisheh
2.  F. Behjati-Ardakani
3.  R. Ebrahimpour
  Status:   Published
  Journal: Applied Soft Computing
  Vol.:  13
  Year:  2013
  Pages:   4570-4578
  Supported by:  IPM
  Abstract:
We present a new classifier fusion method to combine soft-level classifiers with a new approach, whichcan be considered as a generalized decision templates method. Previous combining methods based ondecision templates employ a single prototype for each class, but this global point of view mostly failsto properly represent the decision space. This drawback extremely affects the classification rate in suchcases: insufficient number of training samples, island-shaped decision space distribution, and classeswith highly overlapped decision spaces. To better represent the decision space, we utilize a prototypeselection method to obtain a set of local decision prototypes for each class. Afterward, to determine theclass of a test pattern, its decision profile is computed and then compared to all decision prototypes. Inother words, for each class, the larger the numbers of decision prototypes near to the decision profile ofa given pattern, the higher the chance for that class. The efficiency of our proposed method is evaluatedover some well-known classification datasets suggesting superiority of our method in comparison withother proposed techniques.

Download TeX format
back to top
scroll left or right