Sistema de recomendação para triagem de desordens dos sons da fala infantil baseado em um modelo de consciência de situação

• Ano de defesa: 2018
• Autor(a) principal: Franciscatto, Maria Helena
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Ciência da Computação; UFSM; Programa de Pós-Graduação em Ciência da Computação; Centro de Tecnologia
• Palavras-chave em Português: Consciência de situação; Raciocínio baseado em casos; Teoria do desenvolvimento cognitivo; Sistemas de recomendação; Desordens dos sons da fala; Reconhecimento de fala; Situation awareness; Case-based reasoning; Cognitive development theory; Speech sound disorders; Speech recognition; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/15886

Situation Awareness (SA) involves the correct interpretation of situations, allowing systems to respond to the observed environment and providing support for decision making in many systems domains. Speech therapy is an example of domain where situation awareness can provide benefits, since practitioners should monitor the patient in order to perform therapeutic actions. Also, Case-Based Reasoning (CBR) may contribute to the area, since it uses prior knowledge to solve problems. However, few proposals in the area perform interpretation of a situation to aid in these tasks, just as the CBR methodology is little addressed. For this reason, a recommendation system based on a situation awareness model is proposed, which uses CBR to make useful recommendations to the Speech-Language Pathologists (SLPs) based on the knowledge acquired about the patient. The recommendation system aims to aid clinical decision making, and it is described through an architecture composed of five distinct modules, of which three are specified in the present dissertation - Perpection, Classification and Comprehension. In addition to being based on Situation Awareness and Case-Based Reasoning, the architecture incorporates concepts from the Cognitive Development Theory, as a way of cognitively increasing the proposal and making recommendations that are useful to the SLPs. In the tests performed with respect to the Classification Module, the system was able to classify infant pronunciations with an accuracy above 92%. In the Comprehension Module, error patterns in infant pronunciation were identified, allowing the recommendation of repair strategies to the SLPs and providing support in the screening and therapeutic planning processes.