Academic
Publications
Controlling the diversity in classifier ensembles through a measure of agreement

Controlling the diversity in classifier ensembles through a measure of agreement,10.1016/j.patcog.2005.02.012,Pattern Recognition,Héla Zouari,Laurent

Controlling the diversity in classifier ensembles through a measure of agreement   (Citations: 8)
BibTex | RIS | RefWorks Download
In this paper, a simulation method is proposed to generate a set of classifier outputs with specified individual accuracies and fixed pairwise agreement. A diversity measure (kappa) is used to control the agreement among classifiers for building the classifier teams. The generated team outputs can be used to study the behaviour of class-type combination methods such as voting rules over multiple dependent classifiers.
Journal: Pattern Recognition - PR , vol. 38, no. 11, pp. 2195-2199, 2005
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...Measures to evaluate the diversity of a population play an important role and are studied (or just used) in a variety of areas such as evolutionary algorithms [1]–[16]; swarm intelligence [17]– [20]; numerical integration, uniform design and quasi Monte Carlo methods [21]–[24]; classifier systems [25]–[27], biology, environmental sciences and econometrics [28]–[30]...
    • ...An accurate diversity measure is also particularly important when some decisions are influenced by the diversity via feedback, for instance, when parameters or the structure of an algorithm change (in terms of measured diversity) during the execution [1], [2], [7], [17], [18], [25]...

    Bakir Lacevicet al. On population diversity measures in Euclidean space

    • ...Evidence indicates that diversity within an ensemble is vital for its success [16]-[17]...
    • ...This matrix provides a quantitative performance representation for each classifier in terms of class recognition [17]...

    Cinthia Obladen de Almendra Freitaset al. Distance-based Disagreement Classifiers Combination

    • ...The study results of [19-22] can be used for reference for multi-step choosing system...

    Zhizhou Konget al. Advances of Research in Fuzzy Integral for Classifiers' fusion

Sort by: