• DocumentCode
    3379602
  • Title

    A hybrid approach to object library classification and retrieval

  • Author

    Chang, Chao-Tsun ; Liu, Chung-Shyan

  • Author_Institution
    Dept. of General Sci., Air Commun. & Electron. Sch., Kaohsiung, Taiwan
  • fYear
    1995
  • fDate
    9-11 Aug 1995
  • Firstpage
    278
  • Lastpage
    283
  • Abstract
    A hybrid approach, which combines a faceted scheme and free test analysis, is proposed for classifying and retrieving reusable software components. In our approach, the facet scheme is extended by associating a set of index words, instead of just one, with each facet. The index words are extracted using free test analysis and are ranked by their significance. We also conducted two simple experiments to evaluate the effectiveness of our approach and to determine the factors that contribute to the effectiveness of retrieval. We also found that, free test analysis may still introduce noise, even after two cutoffs are used to remove irrelevant index words. Thus, it may be necessary that, after free test analysis, a filtering phase should be performed for improving retrieval effectiveness
  • Keywords
    filtering theory; libraries; noise; object-oriented programming; software libraries; software reusability; cutoffs; faceted scheme; filtering phase; free test analysis; hybrid approach; index word significance; index words; irrelevant index word removal; noise; object library classification; object library retrieval; retrieval effectiveness; reusable software components; Chaotic communication; Filtering; Frequency; Libraries; Productivity; Science - general; Software performance; Software reusability; Text analysis; Vocabulary;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer Software and Applications Conference, 1995. COMPSAC 95. Proceedings., Nineteenth Annual International
  • Conference_Location
    Dallas, TX
  • ISSN
    0730-3157
  • Print_ISBN
    0-8186-7119-X
  • Type

    conf

  • DOI
    10.1109/CMPSAC.1995.524791
  • Filename
    524791