ABSTRACT
Givenill-behavedpsychologicaldatathatareunlikelytosatisfymetricaxioms,theuseofencodinginsymbolicdynamics,andhenceleadinginto Markovanalyses,isexplored.Variousmeasuresofentropyarecalculated. Thetractabilityofentropicmeasuresforcategorizingthetrajectoriesof nonlineardynamicsthatmaybepresentandchaoticisconsidered,witha focusonthecasewheretherearetwoattractorsandatleastoneheteroclinic orbitbetweenthem.Fast/slowdynamicsaretreatedasaspecialcase.The problemofidenti‡cationisinothercontextstheproblemofdiagnosisin time-varyingpathologies.Somerealdata,selectedfortheirpsychological relevanceinclinical,forensic,andpsychophysicalprocesses,thatareapparently edge-of-chaos and nonstationary, are for comparison analyzed both as metric and discrete and in symbolic encoding.
