Koen Kerremans

Towards Ontology-based E-mail Fraud Detection

The European FF POIROT project (IST-2001-38248) aims at developing applications for tackling financial fraud, using formal ontological repositories as well as multilingual terminological resources. In this article, we want to focus on the development cycle towards an application recognizing several types of e-mail fraud, such as phishing, Nigerian advance fee fraud and lottery scam. The development cycle covers four tracks of development - language engineering, terminology engineering, knowledge engineering and system engineering. These development tracks are preceded by a problem determination phase and followed by a deployment phase. Each development track is supported by a methodology. All methodologies and phases in the development cycle will be discussed in detail

Engineering an Ontology of Financial Securities Fraud

This paper discusses the approach of ontology-based knowledge engineering in FF POIROT, a project to explore the use of ontology technology in information systems against financial fraud. A fraud forensic ontology is being developed from laws, regulations and cases about illegal solicitation of financial products on the web. The knowledge development is based on the DOGMA ontology paradigm, and the derived ontology engineering methodology AKEM. The regulatory ontology engineering is a multi-disciplinary and distributed team work through a series of tasks and deliverables with emphasis on the traceability of decision making in the development. The machine ontology extraction and a manually constructed bilingual terminological database, is used to support the ontology modelling process.

Facilitating ontology (re)use by means of a categorization framework

Ontologies as means for conceptualizing and structuring domain knowledge within a community of interest are seen as a key to realize the Semantic Web vision However, the decentralized nature of the Web makes achieving consensus across communities difficult, thus, hampering efficient knowledge sharing between them To address this problem of heterogeneity we propose a Categorization Framework (CF) that makes it possible to use (multilingual) terminology to specify concepts and concept relations in domain ontologies Such CF could describe the meaning of concepts and concept relations by means of terminological information and external references We believe that such (multilingual) ontology description could enhance the (re)usability and facilitate the coordination of domain ontologies.

PROFILE COMPILER: Ontology-based, community-grounded, multilingual online services to support collaborative decision making

In todaypsilas industries, ontologies and the modern semantic technologies are offered as the potential solutions to support mutual understanding in the collaborative settings. Currently, organizations are unable to cope with collaborative decision making in a particular domain. Mastering domain knowledge and sharing decision rules is a complex yet important issue. In this paper, we first design a generic framework for community-grounded, collaborative decision making enabled, ontology based systems. This framework is constructed with three main engineering tracks: the linguistic engineering track, the ontology engineering track and the group decision engineering track. Each engineering track has its own methodology supported by a tool. Then, we use Semantic Decision Tables to elaborate the knowledge blocks resulting from these three engineering tracks. Semantic Decision Tables are mainly used (1) to elaborate concepts of the domain ontologies, (2) to gather semantically grounded decision rules within a decision group and (3) to provide tabular reports to the users. Finally, an online multilingual knowledge platform - the PROFILE COMPILER is developed as the result. It allows small or medium-sized enterprises (SMEs) to retrieve and personalize existing domain ontology, design and share their decision rules. We illustrate in the domain of Human Resource Management (HRM).

Asking the Right Question: How to Transform Multilingual Unstructured Data to Query Semantic Databases

Ontology engineers have long tried to develop mechanisms to automatically transform natural language statements into queries knowledge-systems can deal with. This has been an enormous challenge as natural languages are highly ambiguous and contexts for disambiguating are seldom identifiable through simple linguistic patterns. To circumvent these difficulties, developers of knowledge bases have often opted for the use of a restricted vocabulary and syntax. Normal users, nevertheless, prefer to express themselves in their language. Special languages or schemas tend to reflect one language – the developer’s – and make extensibility more difficult. Also multilingual access can be more difficult to handle in that way. In this article we present strategies for transforming queries of natural languages into language-neutral representations that can be more easily transformed into semantic queries. We describe a tool that combines a multilingual database and natural processing modules with a semantic database in order to transform queries in Dutch, French and English into queries from which ambiguity at syntactic and semantic levels have been reduced. We focus on certain aspects of natural language such as negation and collocation preferences to deal with semantics.

A Methodological Framework for Ontology and Multilingual Termontological Database Co-evolution

Ontologies and Multilingual Termontology Bases (MTB) are two knowledge artifacts with different characteristics and different purposes. Ontologies are used to formally capture a shared view of the world to solve particular interoperability and reasoning tasks. MTBs are general, contain fewer types of relations and their purposes are to relate several term labels within and across different languages to categories. For regions in which the multilingual aspect is vital, not only does one need an ontology for interoperability, the concepts in that ontology need to be comprehensible for everyone whose native tongue is one of the principal languages of that region. Multilinguality provides also a powerful mechanism to perform ontology mapping, content annotation, multilingual querying, etc. We intend to meet these challenges by linking both methods for constructing ontologies and MTBs, creating a virtuous cycle. In this paper, we present our method and tool for ontology and MTB co-evolution.