Open AccessMolecular descriptors of neural networks with chemical significance
Author(s) -
Sourav Mondal,
Nilanjan De,
Arnab Pal
Publication year - 2021
Publication title -
revue roumaine de chimie
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.135
H-Index - 21
ISSN - 0035-3930
DOI - 10.33224/rrch.2020.65.11.08
Subject(s) - probabilistic logic , artificial neural network , polynomial , degree (music) , variety (cybernetics) , graphical model , property (philosophy) , graph , computer science , mathematics , theoretical computer science , artificial intelligence , physics , mathematical analysis , philosophy , epistemology , acoustics
The probabilistic neural networks (PNNs) are now being analysed to fix a variety of challenges in the diverse fields of science and technology. In chemical graph theory, there are several tools, such as polynomials, functions, etc. that can be used to characterize different network properties. The neighborhood M-polynomial (NM) is one of those that yields neighborhood degree sum based topological indices in a manner that is less time consuming than the usual approach. In this work, the NM-polynomial of 3-layered and 4-layered probabilistic neural networks are derived. Further, some neighborhood degree sum based topological indices are computed from those polynomials. Applications of the present work are interpreted by investigating the chemical importance of the indices. Some structure property models are derived. The graphical representations of the results are also reported.