Automatic summarization assessment through a combination of semantic and syntactic information for intelligent educational systems

Asad Abdi, Norisma Idris, Rasim M. Alguliev, Ramiz M. Aliguliyev

Research output: Contribution to journalArticleAcademicpeer-review

22 Citations (Scopus)

Abstract

Summary writing is a process for creating a short version of a source text. It can be used as a measure of understanding. As grading students’ summaries is a very time-consuming task, computer-assisted assessment can help teachers perform the grading more effectively. Several techniques, such as BLEU, ROUGE, N-gram co-occurrence, Latent Semantic Analysis (LSA), LSA_Ngram and LSA_ERB, have been proposed to support the automatic assessment of students’ summaries. Since these techniques are more suitable for long texts, their performance is not satisfactory for the evaluation of short summaries. This paper proposes a specialized method that works well in assessing short summaries. Our proposed method integrates the semantic relations between words, and their syntactic composition. As a result, the proposed method is able to obtain high accuracy and improve the performance compared with the current techniques. Experiments have displayed that it is to be preferred over the existing techniques. A summary evaluation system based on the proposed method has also been developed.
Original languageEnglish
Pages (from-to)340-358
JournalInformation processing & management
Volume51
Issue number4
DOIs
Publication statusPublished - 2015
Externally publishedYes

Keywords

  • Automatic summary assessment
  • Content coverage
  • Natural language processing
  • Automatic grading
  • Intelligent tutoring systems

Fingerprint Dive into the research topics of 'Automatic summarization assessment through a combination of semantic and syntactic information for intelligent educational systems'. Together they form a unique fingerprint.

Cite this