Evolva: A Comprehensive Approach to Ontology Evolution

Abstract

Abstract. Ontology evolution is increasingly gaining momentum in the area of Semantic Web research. Current approaches target the evolution in terms of either content, or change management, without covering both aspects in the same framework. Moreover, they are slowed down as they heavily rely on user input. We tackle the aforementioned issues by proposing Evolva, a comprehensive ontology evolution framework, which handles a complete ontology evolution cycle, and makes use of background knowledge for decreasing user input. Problem and MethodologyOntologies form the basis of Semantic Web systems. As such, they need to be kept up-to-date for the dependent systems to remain usable. With the increase of complexity and changes occurring in the represented domains, ontology evolution becomes a painstaking and time-consuming process. Thus research has witnessed an increased interest in ontology evolution. We regard ontology evolution as the "timely adaptation of an ontology to the arisen changes and the consistent management of these changes" [10]. "Timely adaptation" suggests a quick adaptation, that can only be achieved by decreasing user involvement in the evolution process. However, most of current approaches heavily rely on user input. Moreover, the definition suggests that a successful evolution can only be achieved by having both "adaptation" and "change management". Yet, as we discuss in Section 4, no existing approach handles the two tasks in one framework. One set of approaches considers evolution as the management of changes performed by users for preserving consistency [11,13,16,17], while another set targets techniques for integrating new knowledge into the ontology [2,5,12,14], without an extensive handling of change and evolution management. Our research tackles the following two main questions:Question 1: How to cover a complete ontology evolution cycle? The focus here is on providing both ontology adaptation and change management.

Adventures of Categories: Modelling the Evolution of Categories During Scientific Investigation

“…Several attempts have been made to conceptualize and structure ontology evolution into a process model that describes the various tasks involved, both to provide a complete understanding of various components of ontology evolution and also as a design requirement for software frameworks to support ontology evolution [14]. Some examples of projects that support the process of ontology evolution are KAON [15], the change-management plugin and PROMPT plugin to Protégé [16], OntoView [17] and the Evolva plugin to the NeON toolkit [18]. KAON supports the whole ontology evolution cycle, from discovering changes to updating ontologies and propagating changes to the dependent artifacts; however, it considers versioning as a separate task and only records the latest conceptualization.…”

Pragmatic Ontology Evolution: Reconciling User Requirements and Application Performance

“…In this phase, we consider the task of selecting a number of concepts to update an ontology as a ranking problem, coherently with the state of the art (e.g., [11,14,15,16]). We thus want to assign a weight to every concept and then update the input ontology with the first n concepts that comply with the requirements.…”

“…The DINO Framework [16], assesses the relevance of a set of candidate triples according to their Levenshtein distance from a set of wanted or unwanted words, specified by domain experts. The Evolva framework [11] measures the relevance of a statement by generating its ontological context from a set of online ontologies and comparing it to the evolving ontology. The DINAMO-MAS system [26] assesses relationships between terms by means of a confidence score that takes in consideration their lexico-syntactic patterns.…”

“…This process is called ontology evolution and it is a critical part of the ontology lifecycle. While the literature proposes a variety of frameworks for ontology evolution [8,9,10,11], essentially most agree on three fundamental steps in the process: i) detection of the need for the evolution, ii) identification of candidate changes, and iii) validation and assessment of these changes, to ensure that the resulting ontology satisfies the given needs.…”

“…Most of the methods for selecting the concepts to be included in an evolving ontology address this task by ranking concepts according to a weight derived from information retrieval metrics [14,15], list of words [16], or online ontologies [11]. These solutions have the advantage of being generic, but present two significant limitations: i) they do not assess the impact of the new version of the ontology on the performance of the relevant applications, and ii) they ignore concept synergy, by weighting the relevance of single concepts rather than the overall impact of a combination of concepts.…”