Loading...
Paper Type
ERF
Paper Number
1530
Description
The paper presents a novelty concept of neural maps with floating nodes (FLOPM), dedicated for classification and prediction tasks. Enhancements of the classic Kohonen neural networks are proposed that allow for the prediction of selected key features, and its justification. The novelty of the approach lies in the concept of floating nodes, i.e. nodes on the Kohonen maps (SOM) that can have any real coordinates (not only discrete values) and can move (float) during the training stage. In future research we would like to test the model thoroughly on various datasets to assess its applicability and compare its efficiency with other data mining tools.
Recommended Citation
Morajda, Janusz and Paliwoda-Pękosz, Grażyna, "A Concept of FLOPM: Neural Maps with Floating Nodes for Classification and Prediction" (2021). AMCIS 2021 Proceedings. 16.
https://aisel.aisnet.org/amcis2021/art_intel_sem_tech_intelligent_systems/art_intel_sem_tech_intelligent_systems/16
A Concept of FLOPM: Neural Maps with Floating Nodes for Classification and Prediction
The paper presents a novelty concept of neural maps with floating nodes (FLOPM), dedicated for classification and prediction tasks. Enhancements of the classic Kohonen neural networks are proposed that allow for the prediction of selected key features, and its justification. The novelty of the approach lies in the concept of floating nodes, i.e. nodes on the Kohonen maps (SOM) that can have any real coordinates (not only discrete values) and can move (float) during the training stage. In future research we would like to test the model thoroughly on various datasets to assess its applicability and compare its efficiency with other data mining tools.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.