Construção e avaliação de um sistema de apoio na correção de mapeamentos conceituais para resolução de casos clínicos

• Ano de defesa: 2016
• Autor(a) principal: Duarte, Jose Marcio [UNIFESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Paulo (UNIFESP)
• Palavras-chave em Português: Natural language processing; Concept mapping; Medical informatics; Processamento de linguagem natural; Mapeamento conceitual; Informática médica
• Link de acesso: https://sucupira.capes.gov.br/sucupira/public/consultas/coleta/trabalhoConclusao/viewTrabalhoConclusao.jsf?popup=true&id_trabalho=4948737; http://repositorio.unifesp.br/handle/11600/47356

Objective: The aim of this research was developing and evaluating a web application based on Natural Language Processing techniques (NLP) that assists professors in correcting Concept Mapping elaborated at the CmapTools® arising the resolution of the health clinical cases. Method: This is a descriptive study of technological development by understanding the process of creating a tool and subsequent evaluation process of their capacity to perform the desired. The study is quantitative consisting in two steps. The first step was a description of development experience automatic corrector based on NLP. The second step consisted the accuracy study. The application web (SAMC ? casos clínicos) accuracy was evaluated based on the comparison of the corrections Concept Mappings developed by students: manual correction performed by the professor and automatic correction performed by the web application with the same evaluation criteria. Results: Presentation of the web application called SAMC ? casos clínicos constituted by Module 1 and Module 2. The Module 1 is responsible for health clinical cases preparation and the Module 2 is responsible for quantitative correction of student's Concept Mapping based on health clinical cases. Also is shown the accuracy evaluation and precision of the developed models that assists NLP tasks at Module 2. Accomplishment a Pilot Test and SAMC ? casos clínicos accuracy evaluation. Conclusion: The SAMC ? casos clínicos proves to be appropriate in supporting and improvement the Concept Mapping teaching strategy to clinical cases resolution with oncology content. The practicality offered by the Module 1 and the automatic correction (Module 2) of Concept Mapping of students reduces stress, errors and time spent by professors.