We need knowledge, not data. The first step is a tool that can read and comprehend texts as well as recognize relevant connections. Along with Prof. Dr. Gerald Petz and his team we collaborated on the project “Named Entity Linking”, taking automated textual analysis to a new level. For a few years now, we have enjoyed our successful close cooperation with the Technical College Steyr (Digital Business – Faculty of Management).
This research project dealt with the semantically correct linking of texts to a knowledge base. For many word processing tasks, entity recognition is a key starting point. “Named entity recognition“ refers to a process that classifies the terms of a text into specific categories. In addition to categorizing the different types of entities occurring in a text, it is crucial as well how they can be semantically linked (“named entity linking“). The purpose of named entity linking is linking the terms of a text to a knowledge base.
Usually the knowledge base Wikipedia is used to link entities. Even though Wikipedia has millions of articles at its disposal, it does not work for more specific domains and contexts. Therefore, our goal was to develop an intelligent named entity linking system which incorporates various knowledge bases and user feedback.
We are very pleased that Tomo, the fastest and most precise text analysis tool in the German-speaking world, can make such good use of these research findings! For more information please consult the following website: FH OÖ Studienbetriebs GmbH.
Prof. Dr. Gerald Petz, Head of Studies, Professor for E-Business