Narrow Results

In this paper we propose an automatic method for building domain ontologies where we leverage the emerging vocabulary from social tagging systems, and the existing semantics in the Linked Open Data cloud to enrich semantically the terms that shape the domain ontology. We systematically capture a domain vocabulary by searching for relevant resources in the folksonomy graph using a spreading activation strategy. We use the vocabulary to identify domain classes and relationships among them by querying knowledge bases published as linked data. We present a case study in the financial domain where we experiment with different settings and show the feasibility of our approach using real folksonomy data.

With the exponential increase of data on the web, the manual acquisition of ontology has become a time-consuming and tedious task. Thus, switching to ontology learning enables the ontologies’ acquisition automation. In this paper, we deal with the phase of concepts’ extraction. Our motivation is to capitalise on the confirmed advantages of deep learning (DL) models and word embedding techniques to automatically extract relevant concepts from large amounts of textual data. A four phases approach is proposed where different models and techniques are applied and a comparative study is achieved: the preprocessing phase, the classification phase, based on DL models, the terms filtering phase, where we experimented and compared three methods to extract the relevant terms, and the semantic enrichment phase experimenting and comparing word embedding techniques to semantically enrich the discovered concepts. The approach is implemented and evaluated on different medical datasets. The obtained results proved the suitability of the experimented models and techniques for the concepts’ extraction.

Semantic web services (SWS) have recently attracted much attention because of their potential to automate all common service tasks such as discovery, composition, invocation and execution. The successful implementation of SWS is profoundly based on the availability of appropriate methods for SWS description. There is, however, no consensus in the SWS arena on the significant service elements that make SWS description comprehensive. Furthermore, semantic annotation of those elements is still a manual process where human involvement is a must. Therefore, automating the annotation process is highly desired as the manual annotation is a difficult, costly and time-consuming process. Few approaches exist aiming to semi-automate the annotation task. These approaches have limitations preventing their wide adoption by the research and practice communities. This paper contributes by analysing and comparing those existing semi-automatic Web service annotation approaches in order to highlight the major challenges and provide fruitful guidelines to aid future research. Particularly, a synthesis of service elements that require semantic annotation is provided and then used to assess the completeness of semi-automated approaches. The comparison leads to significant implications for theory and practice.

Learning Non-Taxonomic Relationships is a sub-field of Ontology Learning that aims at automating the extraction of these relationships from text. This article discusses the problem of Learning Non-Taxonomic Relationships of ontologies and proposes a generic process for approaching it. Some techniques representing the state of the art of this field are discussed along with their advantages and limitations. Finally, a framework for Learning Non-Taxonomic Relationships being developed by the authors is briefly discussed. This framework intends to be a customizable solution to reach good effectiveness in the process of extraction of non-taxonomic relationships according to the characteristics of the corpus.

Along with the rapidly growing scale of relational database (RDB), how to construct domain-related ontologies from various databases effectively and efficiently has been a bottleneck of the ontology-based integration. The traditional methods for constructing ontology from RDB are mainly based on the manual mapping and transformation, which not only requires a lot of human experience but also easily leads to the semantic loss during the transformation. Ontology learning from RDB is a new paradigm to (semi-)automatically construct ontologies from RDB by borrowing the techniques of machine learning, it provides potential opportunities for integrating heterogeneous data from various data sources efficiently. This paper surveys the recent methods and tools of the ontology learning from RDB, and highlights the potential opportunities and challenges of using ontology learning in semantic information integration. Initially, the previous surveys on the topic of the ontology-based integration and ontology learning were summarized, and then the limitations of previous surveys were identified and analyzed. Furthermore, the methods and techniques of ontology learning from RDB were investigated by classifying into three categories: reverse engineering, mapping, and machine learning. Accordingly, the opportunities and possibility of using ontology learning from RDB in semantic information integration were discussed based on the mapping results between the bottlenecks of ontology-based integration and the features of ontology learning.^a

This paper is an extension of our prior work [C. Ma and B. Molnár, Use of ontology learning in information system integration: A literature survey, in Intelligent Information and Database Systems, eds. P. Sitek, M. Pietranik, M. Krótkiewicz and C. Srinilta, Communications in Computer and Information Science, Vol. 1178 (Springer, Singapore, 2020), pp. 342–353]