An ontology-improved vector space model for semantic retrieval

Abstract

Purpose The purpose of this paper is to provide an integrated semantic information retrieval (IR) solution based on an ontology-improved vector space model for situations where a digital collection is established or curated. It aims to create a retrieval approach which could return the results by meanings rather than by keywords. Design/methodology/approach In this paper, the authors propose a semantic term frequency algorithm to create a semantic vector space model (SeVSM) based on ontology. To support the calculation, a multi-branches tree model is created to represent the ontology and a set of algorithms is developed to operate it. Then, a semantic ontology-based IR system based on the SeVSM model is designed and developed to verify the effectiveness of the proposed model. Findings The experimental study using 30 queries from 15 different domains confirms the effectiveness of the SeVSM and the usability of the proposed system. The results demonstrate that the proposed model and system can be a significant exploration to enhance IR in specific domains, such as a digital library and e-commerce. Originality/value This research not only creates a semantic retrieval model, but also provides the application approach via designing and developing a semantic retrieval system based on the model. Comparing with most of the current related research, the proposed research studies the whole process of realizing a semantic retrieval.

A Semantic Matching Method of E-Government Information Resources Knowledge Fusion Service Driven by User Decisions

“…Traditional retrieval systems mainly extract keywords from query sentences and background documents for matching quickly and simply. However, this method ignores the situational awareness of user decision-making and does not uniformly organize the background documents of the system, so retrieval accuracy is relatively low and cannot satisfy the demands of user decision-making activities (Tang et al, 2020). Establishing the user interest model depends on collecting context-aware information, such as the user's personal information and query records, and summarising and dividing the user query vocabulary via ontology theory.…”

A knowledge graph of interlinking digital records: the case of the 1997 Korean financial crisis

“…Wikidata, designed to interact with Wikipedia, is another significant example of the knowledge graph (Vrandečić and Krötzsch, 2014). Recent research on knowledge graphs has focused on frameworks for building knowledge graphs (Ryen et al , 2022), discovery and quality evaluation of knowledge graphs (Tang et al , 2020) and the use of large-scale language models (Alam et al , 2022; Omar et al , 2023).…”