Banca de QUALIFICAÇÃO: LUCAS PEREIRA BRANCO

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : LUCAS PEREIRA BRANCO
DATE: 17/03/2025
TIME: 10:00
LOCAL: Virtual
TITLE:

Dyslexia Diagnosis: an artificial neural networks approach

KEY WORDS:

Dyslexia; Psychometrics; Diagnosis; Artificial Neural Networks; Artificial Intelligence

 

PAGES: 64
BIG AREA: Ciências Humanas
AREA: Psicologia
SUMMARY:

Dyslexia is a neurodevelopmental disorder that affects the reading acquisition process and is characterized by a pattern of difficulties in word reading accuracy, fluency, and reading comprehension. The contemporary perspective understands dyslexia as a multifactorial disorder, which means that it is caused by several factors, whether environmental, neurocognitive, or the interaction between them. The neurocognitive influences in dyslexia are multifactorial and involve deficits in linguistic skills, perception processes, working memory, and executive functions. Identifying these factors and how they act helps not only in the development of intervention strategies but also in identifying people who may be at risk of developing problems related to reading. This project aims to verify a predictive model of the diagnosis of dyslexia in people of both genders, aged between seven and seventeen years. Through the R programming language, a backpropagation artificial neural network will be constructed, a complex statistical and computational technique of deep learning, which can be defined as an artificial intelligence model. The independent variables will come from an online gamified instrument, with 32 linguistic exercises. The dependent variable will be the diagnosis of dyslexia (yes or no). The relevance of the methodological approach is due to the fact that artificial neural networks do not have multicollinearity problems and analyze prediction relationships, also from a non-linear perspective. After construction, the network will be able to generate very precise conclusions about how each independent variable is responsible for predicting the diagnosis of dyslexia and, moreover, will be able to make generalizations for new data, assisting the clinical practice of speech therapists and psychologists and professionals working with written language assessment.

COMMITTEE MEMBERS:
Presidente - 2143029 - CINTIA ALVES SALGADO AZONI
Interno - 1134517 - JORGE TARCISIO DA ROCHA FALCAO
Externo à Instituição - RODOLPHO LUIZ ARAÚJO CORTEZ