Notice: Undefined index: linkPowrot in C:\wwwroot\wwwroot\publikacje\publikacje.php on line 1275
Strona główna > Publikacje
Publikacje
Pomoc (F2)
[42512] Artykuł:

Generalized Tree-Like Self-Organizing Neural Networks with Dynamically Defined Neighborhood for Cluster Analysis

Czasopismo: Lecture Notes in Computer Science   Tom: 8468, Strony: 713-725
ISSN:  0302-9743
ISBN:  978-3-319-07175-6
Wydawca:  SPRINGER-VERLAG BERLIN, HEIDELBERGER PLATZ 3, D-14197 BERLIN, GERMANY
Opublikowano: 2014
Seria wydawnicza:  Lecture Notes in Artificial Intelligence
 
  Autorzy / Redaktorzy / Twórcy
Imię i nazwisko Wydział Katedra Procent
udziału		 Liczba
punktów
Marian Bolesław Gorzałczany orcid logoWEAiIKatedra Informatyki, Elektroniki i Elektrotechniki *337.50  
Jakub Piekoszewski orcid logo33.00  
Filip Rudziński orcid logoWEAiIKatedra Informatyki, Elektroniki i Elektrotechniki *337.50  

Grupa MNiSW:  Materiały z konferencji międzynarodowej (zarejestrowane w Web of Science)
Punkty MNiSW: 15
Klasyfikacja Web of Science: Proceedings Paper

Pełny tekstPełny tekst     DOI LogoDOI     Web of Science Logo Web of Science    
Keywords:

generalized self-organizing neural networks with dynamically defined neighborhood  multi-point prototypes of clusters  cluster analysis  unsupervised learning 


Abstract:

The paper presents a generalization of self-organizing neural networks of spanning-tree-like structures and with dynamically defined neighborhood (SONNs with DDN, for short) for complex cluster-analysis problems. Our approach works in a fully-unsupervised way, i.e., it operates on unlabelled data and it does not require to predefine the number of clusters in a given data set. The generalized SONNs with DDN, in the course of learning, are able to disconnect their neuron structures into sub-structures and to reconnect some of them again as well as to adjust the overall number of neurons in the system. These features enable them to detect data clusters of virtually any shape and density including both volumetric ones and thin, shell-like ones. Moreover, the neurons in particular sub-networks create multi-point prototypes of the corresponding clusters. The operation of our approach has been tested using several diversified synthetic data sets and two benchmark data sets yielding very good results.


POWRÓT
Strona główna > Publikacje