Emerging Named Entity Recognition on Retrieval Features in an Affective Computing Corpus

Affective Computing (AC) is a relatively new, dynamic and interdisciplinary research field. Numerous contributions from fields like computer science, psychology, cognitive science, sociology, physiology and medical science have been made. Consequently, it is difficult to track all recently published trends for early insight utilisation in practise or as basis for innovative research. Even if this fact holds true for many other research fields, AC in this respect is stimulating, due to its dynamic and interdisciplinary characteristics. However, Emergent Entities Recognition is a new concept introduced for early detection and prediction of developing professional terminology. Initial software developments have been completed and briefly analysed in general databases (e.g. MEDLINE). Here, we are interested in its evaluation for AC. In this respect, we have created and used a new Benchmark for Emergent Entities recognition specifially for the field of AC and show evaluation results in comparison to state of the art trained named entity recognition models and to a generic corpus (MEDLINE).