scholarly journals Vektorių kvantavimo metodų ir daugiamačių skalių junginys daugiamačiams duomenims vizualizuoti

2009 ◽  
Vol 50 ◽  
pp. 340-346
Author(s):  
Alma Molytė ◽  
Olga Kurasova
Keyword(s):  
Neural Networks ◽  
Data Structure ◽  
Comparative Analysis ◽  
Multidimensional Scaling ◽  
Vector Quantization ◽  
Self Organizing Map ◽  
Neural Gas ◽  
Self Organizing

Darbe pateikiama lyginamoji dviejų vektorių kvantavimo metodų (saviorganizuojančių neuroninių tinklų ir neuroninių dujų) analizė. Neuronai nugalėtojai, kurie gaunami vektorių kvantavimo metodais, yra vizualizuojami daugiamačių skalių metodu. Tirta kvantavimo paklaidos priklausomybė nuo vektorių nugalėtojų skaičiaus. Išsiaiškinta, kuris vektorių kvantavimo metodas yra tinkamesnis jungti su daugiamačių skalių metodu, t. y. vizualizavus neuronus nugalėtojus „atskleidžiama“ analizuojamųduomenų struktūra.Combination of Vector Quantization and Multidimensional ScalingAlma Molytė, Olga Kurasova SummaryIn this paper, we present a comparative analysis of a combination of two vector quantization methods (self-organizing map (SOM) and neural gas (NG)), based on neural networks and multidimensional scaling that is used for visualization of codebook vectors obtained by vector quantization methods. The dependence of neuron-winners, quantization and mapping qualities, and preserving of a data structure in the mapping image are investigated. It is established that the quantization errors of NG are smaller than that of the SOM when the number of neurons-winners is approximately equal. It means that the neural gas is more suitable for vector quantization. The data structure is visible in the mapping image even when the number r of neurons-winners of NG is small enough. If the number r of neurons-winners of the SOM is larger, the data structure is visible, as well.8px;"> 

scholarly journals Selection of number of neurons for vector quantization methods

2008 ◽  
Vol 48 ◽  
Author(s):  
Olga Kurasova ◽  
Alma Molytė
Keyword(s):  
Neural Networks ◽  
Vector Quantization ◽  
Self Organizing Map ◽  
Data Set ◽  
Selection Of ◽  
Self Organizing

In this paper, a strategy of the selection of the neurons number for vector quantization methods has been investigated. Two methods based on neural networks have been analysed: self-organizing map and neuralgas. There is suggested a way under which the number of neurons is selected taken into account the particularity of the analysed data set.

Assessment of Self-Organizing Map artificial neural networks for the classification of sediment quality

2008 ◽  
Vol 34 (6) ◽  
pp. 782-790 ◽  
Author(s):  
Manuel Alvarez-Guerra ◽  
Cristina González-Piñuela ◽  
Ana Andrés ◽  
Berta Galán ◽  
Javier R. Viguri
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Sediment Quality ◽  
Self Organizing Map ◽  
Artificial Neural ◽  
Self Organizing

New Hardware Architecture for Self-Organizing Map Used for Color Vector Quantization

2019 ◽  
Vol 29 (01) ◽  
pp. 2050002
Author(s):  
Khaled Ben Khalifa ◽  
Ahmed Ghazi Blaiech ◽  
Mehdi Abadi ◽  
Mohamed Hedi Bedoui
Keyword(s):  
Vector Quantization ◽  
Recent Literature ◽  
Network Architectures ◽  
Self Organizing Map ◽  
Description Language ◽  
Field Programmable ◽  
Color Vector ◽  
Hardware Description ◽  
Som Network ◽  
Self Organizing

In this paper, we present a new generic architectural approach of a Self-Organizing Map (SOM). The proposed architecture, called the Diagonal-SOM (D-SOM), is described as an Hardware–Description-Language as an intellectual property kernel with easily adjustable parameters.The D-SOM architecture is based on a generic formalism that exploits two levels of the nested parallelism of neurons and connections. This solution is therefore considered as a system based on the cooperation of a distributed set of independent computations. The organization and structure of these calculations process an oriented data flow in order to find a better treatment distribution between different neuroprocessors. To validate the D-SOM architecture, we evaluate the performance of several SOM network architectures after their integration on a Xilinx Virtex-7 Field Programmable Gate Array support. The proposed solution allows the easy adaptation of learning to a large number of SOM topologies without any considerable design effort. [Formula: see text] SOM hardware is validated through FPGA implementation, where temporal performance is almost twice as fast as that obtained in the recent literature. The suggested D-SOM architecture is also validated through simulation on variable-sized SOM networks applied to color vector quantization.

Sign in / Sign up

Export Citation Format

Share Document