Developing an Ontology Evaluation Methodology: Cognitive Measure of Quality

<p>Ontologies are formal specifications of shared conceptualizations of a domain. Important applications of ontologies include distributed knowledge based systems, such as the semantic web, and the evaluation of modelling languages, e.g. for business process or conceptual modelling. These applications require formal ontologies of good quality. In this thesis, we present a multi-method ontology evaluation methodology, which consists of two techniques (sentence verification task and recall) based on principles of cognitive psychology, to test how well a specification of a formal ontology corresponds to the ontology users' conceptualization of a domain. Two experiments were conducted, each evaluating the SUMO ontology and WordNet with an experimental technique, as demonstrations of the multi-method evaluation methodology. We also tested the applicability of the two evaluation techniques by conducting a replication study for each. The replication studies obtained findings that point towards the same direction as the original studies, although no significance was achieved. Overall, the evaluation using the multi-method methodology suggests that neither of the two ontologies we examined is a good specification of the conceptualization of the domain. Both the terminology and the structure of the ontologies, may benefit from improvement.</p>

Developing an Ontology Evaluation Methodology: Cognitive Measure of Quality

<p>Ontologies are formal specifications of shared conceptualizations of a domain. Important applications of ontologies include distributed knowledge based systems, such as the semantic web, and the evaluation of modelling languages, e.g. for business process or conceptual modelling. These applications require formal ontologies of good quality. In this thesis, we present a multi-method ontology evaluation methodology, which consists of two techniques (sentence verification task and recall) based on principles of cognitive psychology, to test how well a specification of a formal ontology corresponds to the ontology users' conceptualization of a domain. Two experiments were conducted, each evaluating the SUMO ontology and WordNet with an experimental technique, as demonstrations of the multi-method evaluation methodology. We also tested the applicability of the two evaluation techniques by conducting a replication study for each. The replication studies obtained findings that point towards the same direction as the original studies, although no significance was achieved. Overall, the evaluation using the multi-method methodology suggests that neither of the two ontologies we examined is a good specification of the conceptualization of the domain. Both the terminology and the structure of the ontologies, may benefit from improvement.</p>

HEURISTICS AND NUMERICAL METHOD FOR NORMALIZING THE EMPIRICAL CONTEXT IN ONTOLOGICAL DATA ANALYSIS

Областью исследования является онтологический анализ данных, заключающийся в построении формальных онтологий на основе эмпирических данных о слабоструктурированных предметных областях. Предметом исследования является нормализация эмпирического V -контекста - нестрогого соответствия «объекты-свойства» - при ограничениях существования свойств. Задача исследования состоит в разработке численного метода, который реализует эвристический подход к нормализации эмпирических контекстов. В работе используются методы теории множеств и бинарных отношений, модели и методы анализа формальных понятий, а также существующая методология применения ограничений существования свойств для построения формальных онтологий. Отличие и новизна предложенного метода заключаются в более эффективной реализации эвристического подхода за счёт представления системы измеряемых свойств - множества фиксируемых у объектов исследуемой предметной области свойств с заданными на нём ограничениями существования - в виде совокупности субструктур, однородных по виду экзистенциального сопряжения свойств-членов. The research field is ontological data analysis, which consists in the construction of formal ontologies based on empirical data on semi-structured subject domains. The subject of the research is the normalization of the empirical V -context - a non-strict correspondence "objects-properties" - with properties existence constraints. The research objective is to develop a numerical method that implements a heuristic approach to the normalization of empirical contexts. The work uses the methods of the theory of sets and binary relations, models and methods of formal concept analysis, as well as the existing methodology for applying the properties existence constraints to construct formal ontologies. The difference and novelty of the proposed method consists in the more efficient implementation of the heuristic approach by representing the system of measured properties - the set of properties fixed in the objects of the studied subject domain with the existence constraints on it - as a set of substructures that are homogeneous in the form of existential relation of member properties.

Multi-faceted Semantic Clustering With Text-derived Phenotypes

Identification of ontology concepts in clinical narrative text enables the creation of phenotype profiles that can be associated with clinical entities, such as patients or drugs. Constructing patient phenotype profiles using formal ontologies enables their analysis via semantic similarity, in turn enabling the use of background knowledge in clustering or classification analyses. However, traditional semantic similarity approaches collapse complex relationships between patient phenotypes into a unitary similarity scores for each pair of patients. Moreover, single scores may be based only on matching terms with the greatest information content (IC), ignoring other dimensions of patient similarity. This process necessarily leads to a loss of information in the resulting representation of patient similarity, and is especially apparent when using very large text-derived and highly multi-morbid phenotype profiles. Moreover, it renders finding a biological explanation for similarity very difficult; the black box problem. In this article, we explore the generation of multiple semantic similarity scores for patients based on different facets of their phenotypic manifestation, which we define through different sub-graphs in the Human Phenotype Ontology. We further present a new methodology for deriving sets of qualitative class descriptions for groups of entities described by ontology terms. Leveraging this strategy to obtain meaningful explanations for our semantic clusters alongside other evaluation techniques, we show that semantic clustering with ontology-derived facets enables the representation, and thus identification of, clinically relevant phenotype relationships not easily recoverable using overall clustering alone. In this way, we demonstrate the potential of faceted semantic clustering for gaining a deeper and more nuanced understanding of text-derived patient phenotypes.

Patient and Graph Embeddings for Predictive Diagnosis of Drug Iatrogenesis

In the context of the IA.TROMED project we intend to develop and evaluate original algorithmic methods that will rely on semantic enrichment of embeddings by combining new deep learning algorithms, such as models founded on transformers, and symbolic artificial intelligence. The documents’ embeddings, the graphs’ embeddings of biomedical concepts, and patients’ embeddings, all of them semantically enriched with aligned formal ontologies and semantic networks, will constitute a layer that will play the role of a queryable and searchable knowledge base that will supply the IA.TROMED’s clinical, predictive, and iatrogenic diagnosis support module.

ONTOLOGY-BASED RULE COMPLIANCE CHECKING FOR SUBSURFACE OBJECTS

This paper presents a model for representing compliance rules related to subsurface objects. Rules expressed in this model can be automatically evaluated (using SHACL or SPARQL) on existing 3D city models expressed in RDF. The main characteristics of the proposed model are (1) its expressiveness, that comes from the use of formal ontologies for representing the rules and the objects they refer to, (2) its integrative nature, given by the interconnection among the proposed ontologies and the connection of these ontologies with CityGML and IFC (in an ontological form), and (3) its multi-geometry aspect. Preliminary results allow to automatically evaluate formally expressed compliance rules for underground objects in a 3D city model, that will considerably ease the task of professionals of the field.

Information Sciences

The information sciences provide tools for deductive reasoning to supplement the classifications made by the data sciences and the explanations made by explanatory models. Formal ontologies provide a unifying framework for organizing definitions, research findings, and theories. One of the primary purposes of a formal ontology is to use deductive reasoning to answer questions submitted to computer. A general or upper oncology is required to integrate more specialized domain ontologies. The Suggested Upper Merged Ontology is particularly helpful because it consists of 20,000 concepts with connections to both WordNet and FrameNet. WordNet is an electronic dictionary while FrameNet captures co-occurrences of words to provide a thematic context in which words occur. Together, WordNet, FrameNet, and the Suggested Upper Merged Ontology provide an integration of three major information science tools.

A Semantic Matching Strategy for Very Large Knowledge Bases Integration

Over the last few decades, data has assumed a central role, becoming one of the most valuable items in society. The exponential increase of several dimensions of data, e.g. volume, velocity, variety, veracity, and value, has led the definition of novel methodologies and techniques to represent, manage, and analyse data. In this context, many efforts have been devoted in data reuse and integration processes based on the semantic web approach. According to this vision, people are encouraged to share their data using standard common formats to allow more accurate interconnection and integration processes. In this article, the authors propose an ontology matching framework using novel combinations of semantic matching techniques to find accurate mappings between formal ontologies schemas. Moreover, an upper-level ontology is used as a semantic bridge. An implementation of the proposed framework is able to retrieve, match, and align ontologies. The framework has been evaluated with the state-of-the-art ontologies in the domain of cultural heritage and its performances have been measured by means of standard measures.

Searching for the Big Pictures

My goal in searching for the big pictures is to discover novel ways of organizing information in psychology that will have both theoretical and practical significance. The first section lists my reasons for writing each of five articles. The second section discusses an additional five articles that integrate advancements in artificial intelligence and cognitive psychology. The following two sections elaborate on my collaboration with ontologists to use formal ontologies to organize psychological knowledge, including the National Institute of Mental Health Research Domain Criteria, for formulating a biological basis for mental illness. I next discuss strategies for writing integrative articles. The following section describes the helpfulness of the integrations for making psychology relevant to a general audience. I conclude with recommendations for creating breadth in doctoral training.

Formal Ontologies in Information Systems Development: A Systematic Review

Computational ontologies are machine-processable structures which represent particular domains of interest. They integrate knowledge which can be used by humans or machines for decision making and problem solving. The main aim of this systematic review is to investigate the role of formal ontologies in information systems development, i.e., how these graphs-based structures can be beneficial during the analysis and design of the information systems. Specific online databases were used to identify studies focused on the interconnections between ontologies and systems engineering. One-hundred eighty-seven studies were found during the first phase of the investigation. Twenty-seven studies were examined after the elimination of duplicate and irrelevant documents. Mind mapping was substantially helpful in organising the basic ideas and in identifying five thematic groups that show the main roles of formal ontologies in information systems development. Formal ontologies are mainly used in the interoperability of information systems, human resource management, domain knowledge representation, the involvement of semantics in unified modelling language (UML)-based modelling, and the management of programming code and documentation. We explain the main ideas in the reviewed studies and suggest possible extensions to this research.