Scholarlink

The study of three groups of patients with chronic actinic dermatitis showed the presence of type 1 uncertainty. In this case the groups are clinically different, but the samples are statistically the same. To solve this problem used an artificial neural network was used in two special modes: chaos of initial weights of all diagnostic parameters and multiple reconfigurations of the neural network. This results clear division of samples, but also a ranking of their features. This made it possible to find the main diagnostic features based on the ranking of average weights <wi> all diagnostic features. As a result, we found the most important features - order parameters (the problem of system synthesis was solved).</wi>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Artificial Neural Networks in Evaluation of Clinical Parameters of Actinic Dermatitis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • P. E. Konnov,
  • V. V. Eskov,
  • T. V. Voronyuk,
  • I. S. Samoilenko

摘要

The study of three groups of patients with chronic actinic dermatitis showed the presence of type 1 uncertainty. In this case the groups are clinically different, but the samples are statistically the same. To solve this problem used an artificial neural network was used in two special modes: chaos of initial weights of all diagnostic parameters and multiple reconfigurations of the neural network. This results clear division of samples, but also a ranking of their features. This made it possible to find the main diagnostic features based on the ranking of average weights all diagnostic features. As a result, we found the most important features - order parameters (the problem of system synthesis was solved).