A Database Model for an Optimised Relevance Ranking System

Abstract

The need for efficient and intelligent search models has never been more pressing than it is in this era of digital
information explosion, as timely data availability is a long-standing challenge in policy-making and analysis for
low-income and developing countries. Conventional relevance ranking techniques often struggle with scalability,
limited adaptation to user intent, and heterogeneous data sources. An adequately modelled database spurs an
efficient relevance ranking algorithm for efficient document retrieval. An inappropriately modelled database
creates errors related to broken links, which hinder users’ queries from retrieving appropriate documents in the
cloud. Hence, this paper is aimed at developing a model for an Optimised Relevance Ranking system for efficient
document retrieval. The defended PhD theses sampled from the University of Ibadan repository between 1999
and 2021 were used for the database model development. A careful selection of database entities was done from
a document retrieval scenario based on four entity selection rules. An Entity-Relationship (E-R) model was developed with respect to THESES, AUTHOR, CITATION, INSTITUTION, ASBT-KEYWORD (Abstract Keyword), THESES-KEYWORD, USER, SEARCH-RESULT, SEARCH-QUERY, RANKING-FEATURE, and FEEDBACK entities. The E-R model was migrated to a Relational Model (R-M), which is the third normal form of the database, using five conversion rules. The E-R and the R-M models were developed using MySQL Workbench. This approach aims to improve retrieval accuracy and guarantees effective query processing in contemporary information systems by focusing on structured document storage, modularity, scalability, and performance.