| Summary: | Maladaptive aggression is a serious, growing, and ill-understood problem for today's society. This is due, at least in part, to a lack of knowledge regarding how economic, social, environmental, and/or psychiatric factors in?uence the incidence of maladaptive aggression at the individual patient level. Standard statistics have teased out the etiological factors that correlate with the incidence of maladaptive aggression in the population as a whole, but have proven
ine?ective at predicting which patients will display maladaptive aggression and which will not. This failing is likely due to the high number of interactions implicated in the development of maladaptive aggression, the heterogeneous nature of maladaptive aggression, a distinct lack of adequate data sets, or some combination thereof. Thus, the most comprehensive data set on maladaptive aggression available to date was examined with a variety of techniques to overcome some of the
di?culties inherent in predicting maladaptive aggression. The techniques employed were: adapted standard statistics, statistical pattern recognition, machine learning, and a suite of novel predictive analysis tools developed during the process of this dissertation. The results of this investigation provide a method capable of illuminating the complex causes and correlates of maladaptive aggression with both expected and unexpected factors implicated by the current data set. Notably,
this method is easily adapted for use with other data sets and a broad range of predictive problems, especially the investigation of mental illnesses.
|
|---|
