Testing of Support Tool for Plagiarism detection

Motivation

Plagiarism detection tools, also known as text-matching software, are expected
to use the state-of-the-art methods to detect plagiarism. Expected output is the
suspicious document, where plagiarized parts are highlighted (showing WHAT
was copied), together with links to potential sources (showing WHERE these
parts come from. To reveal HOW the source was changed, human consideration
is needed.

There are number of systems available for free or as a paid service, some are
available online, some can be downloaded and used locally. Academics around
the globe are naturally interested in question – which of these systems is the
best?

The most methodologically sound comparison used to be done by prof. Debora
Weber-Wulff and her team. In their last test in 2013, 15 tools were compared.
The selection of the tools was done based on previous comparisons. The
researchers used set of documents mostly in English and German.

Since then, there has not been any systematic and methodologically sound
testing of PD tools since 2013.

Goal

European Network for Academic Integrity wants to fill this gap and started to work on preparation of new methodology of testing including brand new dataset containing documents from multiple European languages plagiarized from various sources and obfuscated by various methods. We believe to deliver first results in 2019.

Members of the group

Head:
Tomáš Foltýnek
Members:
Debora Weber-Wulff
Julius Kravjar
Stella Kleanthous
Alla Anohina-Naumeca
Salim Razi
Dita Dlabolová