Information extraction and exploration are crucial tasks in the era of big data and knowledgeintensive applications. Traditional approaches often need help with data and to leverage semantic knowledge effectively. This paper proposes a semantic knowledge-based framework for information extraction and exploration using Simple Protocol and RDF Query Language (SPARQL) available in the Union Catalogue of Gujarat Colleges (GujCat). The framework incorporates semantic technologies and exploits the power of SPARQL queries to extract and explore structured information from diverse data sources. A semantic information retrieval system provides data using rule-based inference from the ontology. Users can query information from a database or any other data source that can be mapped to Resource Description Framework (RDF) using the "SPARQL Protocols as well as the RDF Query Language." The SPARQL standard, created and supported by the World Wide Web Consortium (W3C), enables developers and users to focus on anything they want to know rather than how a database is structured.
properties defining entities. This makes it possible to conduct far richer types of inquiries on the
numerous unorganised resources than would be feasible with only keyword searches [
Union catalogues are essential for allowing information access across various collections
housed by numerous libraries or repositories. Users can more effectively search for and retrieve
materials thanks to these catalogues, which compile metadata from numerous sources. However,
obtaining and analysing data from union catalogues can frequently be difficult due to the
variability of data sources and the absence of standardised forms. Making a semantically based
framework that uses SPARQL for information extraction and exploration can solve this issue. As
a result, there is a need for a semantic knowledge-based framework that can overcome these
limitations and provide robust information extraction capabilities. Although mining is an
intriguing word to use, it is not a good metaphor to describe the overall knowledge discovery
process [
Addressing these challenges requires the development of a semantic knowledge-based framework for information extraction that leverages the power of SPARQL. SPARQL is a query language specifically designed for querying and manipulating data stored in RDF format. RDF is a standard for representing data on the web in a structured manner, where information is expressed as subject-predicate-object triples. SPARQL enables users to extract, explore, and reason over RDF data through powerful, flexible queries. Main key features and aspects of SPARQL include querying RDF data, pattern matching, variable binding, filtering and expressive functions, joins and optional patterns, aggregation and grouping, inferencing and reasoning, protocol for querying. Such a framework should enable the extraction of structured information from unstructured and semi-structured data sources by effectively incorporating semantic knowledge encoded in ontologies and knowledge graphs. It should provide robust querying capabilities, support the integration of heterogeneous data sources, and facilitate information exploration and visualization for improved understanding and decision-making.
Overall, the problem statement is that the need for a comprehensive framework that harnesses semantic knowledge and SPARQL to overcome the limitations of existing approaches, enabling more accurate, context-aware, and efficient information extraction from diverse data sources.
The purpose of the semantic web is to enhance the current web of text-based
documents/metadata with a layer that machines can understand. To accomplish this, specific
requirements must be met. These include automating the processes of creating semantic
annotations, connecting web content with ontologies, and developing and using ontologies for
interaction [
Ontology learning techniques created an ontology by analysing unstructured and
semistructured material. The term was originally used by Maedche and Staab (1999) [
Cunningham et al. (2006) discovered that this is achieved by taking into account the
relationships between concepts and extracting the information using a regulation expression.
This strategy uses the NLP framework and platform to apply linguistic rules for knowledge
construction [
Shah and Jain (2014) observed that machines can now comprehend the meaning of data and
information and make linkages between various entities and concepts to ontologies. The
Semantic Web can facilitate more advanced information search, discovery, and integration across
multiple domains and applications by annotating and linking data with ontologies [
Another study by Wimalasuriya, D.C., Dou, D. (2010), says that creating and using a semantic
knowledge base is necessary for the knowledge-based framework's execution. Advanced
reasoning techniques can be used to analyse and organise this knowledge base to draw new and
intriguing facts from the provided domain data, which end users can then explore in an informed
manner [
Afolabi et al. (2017), explains an ontology-based association rule mining approach for
extracting knowledge from text. The methodology (Afolabi et al., 2017) incorporates keyword
extraction and weighting based on the term frequency method as part of the data collecting and
cleaning process [
Jiomekong & Tiwari (2023), explains that the research aimed to curate an Open Research Knowledge Graph (ORKG) with papers related to ontology learning and define an approach using ORKG as a computer-assisted tool to organize critical insights extracted from research papers. The ORKG was used to document the "epidemiological surveillance systems design and implementation" research problem and prepare related work. open research knowledge grapy as computer assistant tool consist of five tools like knowledge elicitation, Knowledge analysis and interpretation, Template creation, Knowledge representation, Verification and validation. [20].
Amara et al. (2023), provides an in-depth analysis of semantic interoperability in Industry 4.0,
highlighting its core concepts, problems, and implications for intelligent manufacturing. It
explores the potential of semantic technologies like ontologies, linked data, and standard data
models [
Khorashadizadeh et al. (2023), evaluates the use of ChatGPT in GPT-3.5, a large foundation model, to improve knowledge graph construction and completion. The qualitative analysis reveals ChatGPT's potential, but challenges like bias, hallucinations, and high computational costs must be addressed [21].
The research effort reported in this paper claims that domain knowledge can significantly contribute to the activities of Information Extraction from unstructured data. This section presents a proposed process of information extraction as shown in Figure 1 for comprehensive, semantic-driven methodology for utilising that knowledge to improve the machine learning techniques employed in relation extraction.
• • • • • • •
Preprocessing and merging the domain and construction of the knowledge map. Recognising the named entities.
Subject mapping with the DDC schemes.
Collecting the training datasets from structured datasets.
Building rules based semantic extraction using SPARQL.
Extracting Relations from the unstructured data by using the classification models.
In this section, study discuss proposed schema of GCOnto model for searching for semantic
knowledge as shown in Figure 2 and to achieve that researcher installed Apache Jena Fuseki and
used Protege tool and for information extraction, researcher used a sample dataset with
subjectwise metadata available in the Union Catalogue of Gujarat colleges (GujCat) using SPARQL query.
A Semantic-based Framework for Information Extraction and Exploration of a Union Catalogue
using SPARQL involves leveraging RDF and ontologies to represent the catalogue data and using
SPARQL queries for extracting and exploring the information. The Web Ontology Language
(OWL) is often used, assisted by ontology modelling tools like Protégé. [
Ontologies need a well-designed language and rigorous logical reasoning to be an effective
method. Ontological knowledge bases come with the T-box, a terminological formalism, and the
A-box, a declarative formalism. [
Ontology GCOnto = (Terminological Formalism + Assertional Formalism, Graph)
Formulate SPARQL queries to explore the relationships between entities and discover patterns and insights.
}
WHERE { ?book rdf:type :Book . ?book :hasTitle ?title . ?book :hasAuthor ?author . ?book :hasSubject ?subject. ?relatedBook rdf:type :Book . FILTER (?book != ?relatedBook) .
The primary semantic relationship among concepts is inheritance. When querying a parent class, it should encompass all its subclasses or child classes. Figure 3 shows the computer science subject-wise metadata exploration from GCOnto using SPARQL. For instance, when searching subjects related to "Computer Science, Information & General Works", the query statement would be:
FROM GCOnto WHERE { f: Computer science, knowledge and systems f: value ?all_sub_class} <rdfs:Class rdf:ID=”Computer science, information & general works”></rdfs:Class> <rdfs:Class rdf:ID=”Computer science, knowledge and systems”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, information & general works”/> </rdfs:Class> <rdfs:Class rdf:ID=”General Knowledge of Computer”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, knowledge and systems”/> </rdfs:Class> <rdfs:Class rdf:ID=”The book (writing, libraries, and book-related topics)”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, knowledge and systems”/> </rdfs:Class> <rdfs:Class rdf:ID=”Systems”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, knowledge and systems”/> </rdfs:Class> <rdfs:Class rdf:ID=”Data processing & computer science”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, knowledge and systems”/> </rdfs:Class> <rdfs:Class rdf:ID=”Computer programming, programs & data”></rdfs:Class> <rdfs:SubClassOf rdf:resource=”#Computer science, knowledge and systems”/> </rdfs:Class>
• • • •
A semantic-based framework that enables efficient information extraction and exploration from GCOnto using SPARQL queries.
Enhanced semantic interoperability and data integration among heterogeneous sources. Improved query performance and response times, even with large-scale data. A foundation for future research and development in the realm of semantic-based information extraction and exploration.
In this paper, study is reviewing information extraction from unstructured and semistructured data in an accurate and efficient way using a semantic based framework. For semantic based information retrieval, it has proposed an ontology concept for union catalogue and SPARQL query to retrieve the result in an efficient way. Inference based on inheritance relation in subject hierarchy increases the precision and recall of result set data. Generating the semantic content from the proposed framework makes it a more interesting concept. GCOnto system can be used for identifying the semantic content for the semantic web and also implemented Ontology Based web service for results.
While searching the content in ontology, it returns a large number of results. Using the advanced techniques of NLP, analyse the outcome and find the most relevant subset of data to improve the relevancy of search text with the result.
Web. Springer Nature Switzerland. 14382 (2023) 82–96. https://doi.org/10.1007/978-3031-47745-4_7 [20] Jiomekong, A. Tiwari, S. An approach based on Open Research Knowledge Graph for Knowledge Acquisition from scientific papers. (2023) https://doi.org/10.48550/ARXIV.2308.12981 [21] Khorashadizadeh H. Mihindukulasooriya N. Tiwari S. Groppe J. Groppe, S. Exploring InContext Learning Capabilities of Foundation Models for Generating Knowledge Graphs from Text (2023). https://doi.org/10.48550/arXiv.2305.08804