Tag line generating system using information on the web

Hiroaki Yamane, Masafumi Hagiwara

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

This paper proposes a tag line generating systemusing information extracted from the web. Tag lines sometimes attract attention even when they consist of indirect word group of the target. We use web information to extract hidden data and use several tag line corpora to collect a large number of tag lines. First, knowledge related to the input is obtained from the web. Then, the proposed system selects suitable words according to the theme. Also, model tag lines are selected from the corpora using the knowledge. By inserting nouns, verbs and adjectives into model tag lines' structure, candidate sentences are generated. These tag line candidates are selected by the suitability as a sentence using a text N-gram corpus. The subjective experiment measures the quality of system-generated tag lines and some of them are quite comparable to human-made ones.

Original languageEnglish
Pages (from-to)185-193
Number of pages9
JournalJournal of Advanced Computational Intelligence and Intelligent Informatics
Volume17
Issue number2
DOIs
Publication statusPublished - 2013 Mar

Keywords

  • N-gram
  • Sentence generation
  • Tag lines
  • Text mining
  • Web information

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Fingerprint Dive into the research topics of 'Tag line generating system using information on the web'. Together they form a unique fingerprint.

Cite this