European Conference on Machine Learning and Principle and Practice of Knowledge Discovery in Databases (ECML PKDD)

[img_assist|nid=2947|title=|desc=|link=none|align=left|width=100|height=31]The ECML PKDD is a technical conference on cutting-edge innovations in the areas of machine learning and knowledge discovery. It was held in Barcelona, Spain, at the Universitat Autonoma de Barcelona. The keynote by Prof. Hod Lipson offered insights on current trends in robotics and machine learning. Lipson's team aims to create a machine learning algorithm capable of deriving principles governing the natural sciences based on the automated analysis of raw data.

Stefan Gindl, researcher at the Department of New Media Technology, gave a talk on the department's current work on sentiment detection. The presented paper compares "Different Aggregation Strategies for Generically Contextualized Sentiment Lexicons". The underlying technology refines classic sentiment lexicons by integrating context knowledge to flexibly adapt the contained sentiment values. This contribution to the special track of the DyNaK workshop (Dynamic Networks and Knowledge Discovery) presented parameter settings to optimize this process.