A carregar...
Publicação
Contribution of discriminant analysis in the classification of anxiety disorders
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.28 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
In recent decades, the importance given to the treatment of mental disorders (MDs) has increased. As some MDs can be prevented if diagnosis occurs in the early stages of the disease, this thesis aims essentially to revisit some statistical discriminant methods emphasizing their potential role to classify patients into different MD categories. Three groups (nervous, psychotic, and healthy) of fifty individuals each were analyzed with the purpose of best separating the three groups. Factorial discriminant analysis (FDA) was used in an initial set of thirty variables, obtaining a misclassification rate of 4.(6)% in the training set and 10.(6)% in the testing set. Each variable was analyzed over its contribution to the discrimination of the groups under matter in which, succeeding the first FDA performed, the correlation between each variable and each discriminant function (DF) was examined, the partial F-test was made for every variable, and the standardized DF coefficients were analyzed, where fifteen variables were selected to take part in a new analysis. The new FDA model obtained a misclassification apparent rate of 10% and 14.(6)% in the training test and testing test, respectively. This model was compared to a stepwise discriminant method of eighteen variables, where it was obtained a misclassification apparent rate of 10% in the training set and 16% in the testing set. The reduced model is then recommended over the two other models, as it is more adequate for the main purpose of this thesis, as its complexity is lower when compared to the full model, and also outperformed the stepwise discrimination model. In conclusion, FDA has proven itself to be a potential good method to diagnose patients considered in this study.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Information Analysis and Management
Palavras-chave
Mental disorders Nervous Psychotic Principal Component Analysis Factorial Discriminant Analysis SDG 3 - Good health and well-being