A QA-TSK fuzzy model vs evolutionary decision trees towards nonlinear action pattern recognition

A comparison among three linear methodologies, a novel auto-adjusted fuzzy quadruple TSK model (QA-TSK) and two evolutionary decision tree representations, is presented in this paper. The three architectures make use of a vast number of primitives utilised to reconfigure and evolve their internal structures of the classifier models so that to discriminate among spatial physical activities. Such primitives like statistical features employ a twofold role, initially to model the data set in a dimensionality reduction preprocessing and finally to exploit these attributes to recognise pattern actions. The performance statistics are being utilised for remote surveillance within a smart environment incorporating an ubiquitous 3D marker based tracker which acquires the timeseries data streams, whereas activity recognition statistics are being generated through an off-line process.