TIME SERIES ANALYSIS OF INTEREST IN PSYCHIATRYAuthor:
PARRA ARÉVALO MARÍA ISABEL.
Year:
2003.
University:
EXTREMADURA [
www.unex.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: FACULTAD DE CIENCIAS.
Summary: The aim of the dissertation is model and characterize the symptoms of certain psychiatric disorders, valuing the appropriate therapeutic action, in order to have objective means of diagnosis and prediction of the clinical outcome of the patient. Specifically, it is considered a study based on actual data provided by the psychiatric unit at the University Hospital Infanta Cristina de Badajoz, related to the weight daily for patients diagnosed with Behavior Disorders Food and electroencephalographic patient records to which it administers Therapy Electroconvulsiva. To avoid bias in the conclusions taken drinks precautions as using double-blind technique or working with control groups. In both cases, performed an exhaustive study of the corresponding time series, using a common methodology based on specific techniques of nonlinear dynamic models. In order to carry out data analysis, has developed a package of functions (programmed in mathematics) that facilitates the analysis graph up from the calculation of measures of complexity. As most relevant conclusions can be highlighted. * The study of the evolution of weight in patients with eating disorders in the diagram gas can be characterized from the dimension fratal of attractor. * Through dynamic symbolic or fractal dimension, we can estimate the end of the convulsion in Therapy Electroconvulsiva, so that there is a better match with the observations of clinicians to those on other algorithms in use. * Preliminary results indicate that the value which stabilizes the fractal dimension after a seizure is directly linked to the overall improvement of the patient.
