AAU Student Projects - visit Aalborg University's student projects portal
A master's thesis from Aalborg University
Book cover


Improving quality of keyword query results using ESKuSE

Author

Term

4. term

Publication year

2003

Abstract

Denne rapport præsenterer ESKuSE, en hurtig og effektiv nøgleordssøgemaskine til relationelle databaser. Den repræsenterer databaseskemaet—tabellerne og deres relationer—som en graf og bruger denne graf aktivt under kørsel af forespørgsler. Systemet udnytter hints på forskellige niveauer i databasen (fx tabel- og kolonnenavne, relationer og andre metadata) til at lave kvalificerede gæt om data på et konceptuelt niveau, ikke kun på et rent logisk niveau. Denne grafbaserede, hint-drevne tilgang styrer, hvordan forespørgsler udføres.

This report introduces ESKuSE, a fast and efficient keyword search engine for relational databases. It represents the database schema—the tables and how they relate—as a graph and relies heavily on this graph when running queries. The system uses hints found at different levels of the database (for example, table and column names, relationships, and other metadata) to make informed guesses about the data at a conceptual level, not just a purely logical level. This graph-based, hint-driven approach guides how queries are executed.

[This abstract was generated with the help of AI]