Linguistic Classification System

The Linguistic Classification System (LCS) was developed at the University of Nijmegen, The Netherlands, in the course of the EC/IST projects DORO and PEKING. The LCS is a classification system based on machine learning algorithms Rocchio, Balanced Winnow and SVMLight, especially suited for the task of classifying text documents. The LCS allows different document representations (keywords with or without lemmatization, linguistic phrases), hence the L in its name.

The LCS can either be used through a Java-based API or through Web services. While the API supports the training of the categorisation system as well as the categorisation of unseen patents, the Web services currently offers functionality of categorisation. The availability of training through Web services is an objective for a future release. 


How to make your own project visible on the IRF Prototype Portal

All IRF members are invited to present their demos and prototypes (even if early-stage) on the IRF Prototype Portal. This is an excellent opportunity to collect feedback from end-users and increase the visibility of your project. Please contact for more information.