Sit-based functional dependency extraction

Balázs Tusor, János T. Tóth, Annamária R. Várkonyi-Kóczy

Research output: Contribution to journalArticle

Abstract

In the digital age, being able to determine the relationships between attributes in datasets is advantageous for many fields of information technology, such as data mining, machine learning, etc. There are inherent rules, functional dependencies that are derived from the nature of the data, of which many are not trivially obvious to the human data manager. This paper presents a new method to extract such relationships. It uses Sequential Indexing Table structures that can implicitly indicate if the values of an attribute are determined by the combination of the values of the attributes preceding it in the layered architecture, i.e. is functionally dependent on those attributes. A new algorithm is given to use this feature to extract functional dependencies, and the performance is analyzed using real-life datasets.

Original languageEnglish
Pages (from-to)65-81
Number of pages17
JournalActa Polytechnica Hungarica
Volume16
Issue number10
DOIs
Publication statusPublished - Jan 1 2019

    Fingerprint

Keywords

  • Data mining
  • Data processing
  • Functional dependencies
  • Functional dependency extraction
  • Sequential indexing tables

ASJC Scopus subject areas

  • Engineering(all)

Cite this