Concept Extractor - Ein flexibler und domänenspezifischer Web Service zur Beschlagwortung von Texten

We describe a flexible and modular system for keyword extraction and attribution which operates on top of a text mining engine. Texts are analysed in comparison with a large reference corpus and key words are determined using a frequency based method for determining relative term significance. Additionally, selected terms may be expanded using large knowledge bases on inflected forms, orthographic variants, synonyms and multi word terms. This solution is realised as a web-based service which can easily be integrated into existing content management systems.