Author(s): Balázs TUSOR, Ondrej TAKÁČ, Annamária R. VÁRKONYI-KÓCZY, Štefan GUBO
Title: STRICT AND APPROXIMATE FUNCTIONAL DEPENDENCY EXTRACTION WITH SEQUENTIAL INDEXING TABLES-BASED SEARCH TREES
Source: F. Filip, Zs. Gódány, Š. Gubo, E. Korcsmáros, S. Tóbiás Kosár, T. Zsigmond (eds.): 16th International Conference of J. Selye University. Sections of the Faculty of Economics and Informatics. Conference Proceedings
ISBN: 978-80-8122-509-3
DOI: https://doi.org/10.36007/5093.2024.309
Publisher: J. Selye University, Komárno, Slovakia
PY, pages: 2024, 309-320
Published on-line: 2024
Language: en
Abstract: Data relation analysis has been a very important field of data science for the past few decades, in which the goal is to discover relationships between data attributes. Functional dependencies are among the most basic of such relations, defining a strict determination between the values of the attribute set on the determinant side and the values of the attribute on the dependent side of the relation. Approximate functional dependencies allow a certain amount of deviation from strict dependencies, and thus, can indicate relationships that are true for a large extent, with only a few exemptions. In this paper, a new method is presented for the extraction of strict and approximate dependencies, which applies the same base idea as the SFIT, but with the combination of binary search trees to reduce the required memory without sacrificing much of the operational speed.
Keywords: functional dependency extraction, approximate functional dependency extraction, indexing tables
Fulltext (PDF)