Automatic Context-Dependent Text Document Summarization

A. Yu. Chugunov , e-mail: arcady.chugunov@gmail.com

This article deals with the urgent task of context-dependent summarization considering user's search query. It's provided a brief review of existing studies on the subject, an analysis of their advantages and disadvantages. A mathematical model of annotating documents is considered. Special attention is given to the method of spectral estimation of lexical units of text with the appropriate mathematical tools. The results of the software implementation of the KGCD-algorithm on the news articles are provided. Via the test results there made some conclusions and the directions of extending the functionality of the algorithm selected for the implementation are given.

Keywords: automatic summarization, document, context-dependent approach, dynamical summarization, spectral lexeme characteristics, Kenny-Goodman algorithm

pp. 39–46