Full Length Research Paper
Abstract
Artificial neural network (ANN) is defined as computational models with structures derived from the simplified concept of the brain in which a number of nodes are interconnected in a network-like structure. The most used ANNs architecture for pattern recognition and classification is the self-organizing map (SOM). SOM is a powerful visualization tool as it is able to reduce dimensions of projections and displays similarities among objects and was successfully used in several applications with chemistry database. In this work, we used SOM as good methodology of classification of a database containing various types of compounds from theAsteroideae subfamily (Asteraceae). The Kohonen neural network was trained using Matlab version 6.5 with the package Somtoolbox 2.0. Some chemical evolutionary descriptors and the numbers of occurrences of 12 chemical classes in different taxa of the subfamily were used as variables. The study shows that SOM applied to chemical data can contribute to differentiate genera, tribes, and branches of subfamily, as well as to tribal and subfamily classifications ofAsteroideae, exhibiting a high hit percentage comparable to that of an expert performance, and in agreement with the previous tribe classification proposed by Funk.
Key words: Asteraceae, Asteroideae, self-organizing maps, secondary metabolites.
Abbreviation
ANN, Artificial neural network; SOM, self-organizing map; SOM Kohonen map, Kohonen self–organizing feature map; O, oxidation state; S, skeletal specialization; OS,oxidation step (OS for ring A and as OSB for ring B).
Copyright © 2024 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0