Introducing context in syllable based emotion tracking

In this paper, we present a further step in the development of an emotion tracking system based on phonetic syllables and machine learning algorithms. A system built on phonetically defined units has advantages both on the side of the amount of data needed to train the classifier and on the ability of improving our knowledge about how humans use speech to recognize emotions on the base of the physical meaning of each used feature. Since the features extraction frequency is intrinsically variable, however, it is necessary to study how to represent context and dynamics as well as to evaluate the effects of their introduction in the system. The goal of this study is to evaluate the effects of context in a previously presented system working on isolated syllables only. Obtained results show that the system performance is improved.