UDC 007:681.512.2
APPLICATION OF HIERARCHICAL NUMBERS THEORY IN ICF TAXONOMY DESIGN FOR NEURAL NETWORK OPTIMIZATION
I. Yu. Kashirin, Dr. Sc. (Tech.), full professor, department of computational and applied mathematics, RSREU, Ryazan, Russia; orcid.org/0000-0003-1694-7410, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The article considers a new approach to the design of neural networks using ICF-ontologies, structured in accordance with certain transformation rules. The rules are described using the theory of hierarchical numbers. Applied ICF ontology is used to describe the semantic part of input data of training sample based on the calculation of semantic similarity of concepts and features. The model of knowledge, redesigned on the basis of transformational rules, allows us to optimize input data sets for big data mining. Software implementation of the proposed approach is based on learning data analysis algorithms. The experiments performed using Python v.3 (Anaconda 3) toolkit show the effectiveness of formal data conversion apparatus described in the article. The aim of the work is to create a science-intensive technology for designing Data Mining algorithms with training to solve big data analysis problems based on the application of hierarchical numbers theory.
Key words: Data Mining, learning algorithms, transformation rules, ICF taxonomy, hierarchical numbers, semantic proximity, clustering, neural networks