Neuronová síť pro asociace textů

Neural Network for Text Associations

článek ve sborníku mimo WoS a Scopus

en

The paper proposes a new neural network based on the principle of the Restricted Coulomb Energy (RCE) classifier. The network has three layers of neurons and it works as a heteroassociative memory. The short principles of the proposed neural network – topology, learning and retrieving – are described in the paper. Experiments with text lines associations that have been done with this neural network and with the Bidirectional Associative Memory (BAM) and a comparison of these results are described in the paper, too.

Příspěvek představuje neuronovou síť založenou na bázi RCE klasifikátoru – síť má tři vrstvy neuronů a pracuje jako heteroasociativní paměť. Stručně je popsána topologie, učení a odezva sítě, experimenty s textem a jejich srovnání s výsledky dosaženými obousměrnou asociativní pamětí BAM. 

The paper proposes a new neural network based on the principle of the Restricted Coulomb Energy (RCE) classifier. The network has three layers of neurons and it works as a heteroassociative memory. The short principles of the proposed neural network – topology, learning and retrieving – are described in the paper. Experiments with text lines associations that have been done with this neural network and with the Bidirectional Associative Memory (BAM) and a comparison of these results are described in the paper, too.

Neural Network, BAM, RCE Classifier

2001

01.01.2000

Marq software s.r.o.

Ostrava

80-85988-51-8

Proceedings of XXIInd International Colloquium ASIS 2000

145–150

6