Paris13 Terminae

From ONTORULE Show Case

Motivation

In the context of companies relying heavily on policies and business rules, it is often important to trace knowledge to textual sources. Also, while domain knowledge cannot be fully automatically extracted from texts, ontology acquisition from text is an important step in the knowledge acquisition process, since texts, unlike experts, are always available, partly reflect the domain conceptualisation (TBox) and may contain pieces of factual knowledge (ABox). TERMINAE is a platform that assists users in designing termino-ontological resources from texts. It can be used by terminologists to build terminological forms and by knowledge engineers to build either thesaurus expressed in SKOS or ontologies organising concepts and lexical units in a formal way supporting inferences. This platform allows to link textual elements to terminological and conceptual resources. The acquisition corpus may contain one or several documents. The supported languages are English and French.

Terminae Workflow

Description

The Terminae workflow consists in execution of four steps. Each step is performed using a different perspective. The initial inputs are:

- an acquisition corpus which has to be transformed using the term extractor Lingua-YaTeA and the lemmatizer TreeTagger.

- a domain in which the business rules are used.

1. The first step consists in performing linguistics actions as removing irrelevant symbols or terms or clustering similar terms.
2. The second step consists in creating terminological forms. Each terminological form involves the description of a term that may be relevant in the domain. The set of terminological forms describes a terminology of the domain.
3. In the third step, each meaning of each term gives rise to a termino-conceptual entity. The set of termino-conceptual entities describes a termino-conceptual network where links are either generic-specific links or properties between termino-conceptual entities.
4. In the last step, each termino-conceptual entity gives rise to a conceptual entity. The conceptual entities are created using Neon Toolkit editor and each conceptual entity (class, object property) is linked to a termino-conceptual entity using an annotation.

The termino-conceptual network may be saved in a SKOS file and the ontology may be saved in a OWL file holding termino-conceptual annotations.

Availability

Terminae is available in D1.1 deliverable: software and user manual. A new version was delivered in D1.4.