Scholarlink

Abstract <p>The aim of the study is to study the optimization of the transformation of multimodal information based on semantic fields for the space of vector text and visual documents. The research methodology includes the creation of a vector model of genetic selection of semantic fields in the problems of classifying text and visual documents when transforming multimodal information to ensure its accessibility for people with visual impairments. The accuracy of the classifier by nearest neighbors is used as the objective function for genetic optimization. As a basis for the experimental study, we will choose a standardized collection of newsgroup messages 20-Newsgroups. A conclusion is made about the effectiveness of using genetic algorithms to optimize a set of semantic fields that form the basis of the vector space of documents in the classification analysis of text and visual documents when transforming multimodal information to ensure its accessibility for people with visual impairments.</p>

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

Study of Optimization of Transformation of Multimodal Information Based on Semantic Fields for the Space of Vector Text Documents

  • 期刊论文

  • 发表日期：

    2025年6月1日

查看全文
  • A. V. Degtyar’,
  • N. E. Nekrasov,
  • T. A. Markina,
  • I. B. Gosudarev

摘要

Abstract

The aim of the study is to study the optimization of the transformation of multimodal information based on semantic fields for the space of vector text and visual documents. The research methodology includes the creation of a vector model of genetic selection of semantic fields in the problems of classifying text and visual documents when transforming multimodal information to ensure its accessibility for people with visual impairments. The accuracy of the classifier by nearest neighbors is used as the objective function for genetic optimization. As a basis for the experimental study, we will choose a standardized collection of newsgroup messages 20-Newsgroups. A conclusion is made about the effectiveness of using genetic algorithms to optimize a set of semantic fields that form the basis of the vector space of documents in the classification analysis of text and visual documents when transforming multimodal information to ensure its accessibility for people with visual impairments.