An ontology development approach using concept maps driven by automatic term extraction

Abstract

A number of ontology engineering methodologies have been proposed to date. Distinct methodologies rely on different techniques and activities for developing ontologies. The concept mapping technique has been recently used for developing ontologies related to different domains. However, the existing approaches using concept mapping do not make use of any automatic term extraction process to ease the overall ontology development process and their evaluation procedures are not robust to specifically meet the need for ontology designs emerging from concept maps. This paper proposes a new approach which uses the concept mapping technique coupled with a term extraction engine. The engine automatically extracts the candidate terms from the competency questions which reduces ontology development time and efforts. Furthermore, considering the importance of evaluation, a robust evaluation procedure is proposed which validates the ontologies at both the formal level and the graphical level.

A systematic literature review of automatic ontology construction

“…Meanwhile, [19] has done automation in the ontology conceptualization phase, the second phase of the methodology, and [12] in the implementation phase which is the third phase of the methodology. However, [11] and [23] have done automatic construction in more than one phase whereas [11] has done automatic construction in the ontology specification and evaluation phase, while [23] in the ontology specification, conceptualization, and implementation phase. As can be seen from Table 6, most of the papers have done automatic construction in the ontology specification phase by doing term extraction activity.…”

“…As can be seen from Table 6, most of the papers have done automatic construction in the ontology specification phase by doing term extraction activity. Sharef et al [11] has used the engine based on POS tagging coupled with patternbased extraction techniques to automatically extracts the candidate terms from the competency questions, in [17] has used TF-IDF measure to extract relevant concepts from the corpus, in [22] used relevance score based on DP and DC to extract useful terms, and [23] performs the activity of constructor and converter for pre-processing and use TF-IDF for term weighing. Besides, [30] has applied NLP to extract important information from academic papers by using Python 2.7 and NLTK, meanwhile [31] has used a term scoring algorithm for term and phrases extraction from textual corpus to generate the domain terminology.…”

“…To implement the ontology, [12] has added the Celfie plugin to extract data from an Excel sheet to create the ontology, the individuals, data properties, and object properties, and apply the Fact++ reasoning technique to the ontology to gather logical consequences from a set of asserted facts or axioms, and [23] use OWL language to build an ontology. Lastly, [11] also done automation in ontology evaluation by using DL as formal query language and executing using DL-Query Tab to check ontology correctness, using concept map as a reference frame to compare with implemented ontology, and OntoGraf Plugin to capture different snapshots of implemented ontology.…”

“…− Cmap Tools (2013) to construct a concept map [11] − Protégé 5.0.0 software / Protégé IDE [11], [12], [19], [23], [28], [29], [32], [33] − Apache Jena API to write ontology in RDF file [8] [24] − Jambalaya plug-in to supports the visualization of Arabic letters in ontology visualization [19] − Neo4j or Watson to generate KG from text [15] − Java API for WordNet Searching (JAWS) library to obtain synsets from WordNet [22] Evaluation − YAGO and DOLCE as reference ontologies [15] − HAABSA framework in Python to evaluate created ontology [22] − OntoGain tools [8]…”

“…There is no one-size-fits-all approach to developing ontologies. Instead, the idea of combining different methodologies and techniques is supported by practitioners in the field of ontologies [11]. However, the general phases involved in the ontology construction methodology proposed by [12], as shown in Figure 1, can be used as a guideline to construct the ontology.…”