Banca de QUALIFICAÇÃO: PATRÍCIA MAYARA MOURA DA SILVA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : PATRÍCIA MAYARA MOURA DA SILVA
DATE: 22/03/2021
TIME: 14:00
LOCAL: Vídeo Conferência
TITLE:

achine learning applied to gait execution in patients with type 2 diabetes.

KEY WORDS:

Keywords: Diabetes Mellitus; Gait; Machine Leaming.

PAGES: 82
BIG AREA: Ciências da Saúde
AREA: Fisioterapia e Terapia Ocupacional
SUMMARY:

lntroduction: Diabetes is characterized by a set of metabolic diseases that
can cause several changes. One of them is in the sensorimotor function that
generates changes in gait execution, such as longer support phase, shorter
steps, and inadequate plantar pressure distribution. Quantitativa methods for assessing changes in
gait pattems can be decisive in designing treatment strategies. Besides, they can help in
preventing complications caused by diabetes. With advances in machine leaming techniques (ML),
automated pattem recognition in the face of the immense amount of data has become an essential tool
in the medicai field due to the ability to predict clinicai complications before the disease gets
worse. Objectives: To implement ML model on gait assessment data of diabetic patients to predict
clinicai complications of diabetes. Methods: The study will involve two types of methodological
modalities: 1) Elaboration of a protocol and a Systematic Review; 2) Development and improvement of
a predictive ML model that analyzes gait execution relating it to diabetes clinicai complications.
To carry 

out this study, the data will be provided through a partnership with the Florida
lnternational University (FIU) during sandwich doctorate (CAPES/PRINT grant
- No. 02/2020) expected to start in May 2021. These data will be prepared and preprocessed to
further implementation on different ML models. The Systematic Review results performed in this
study will permit choose the ML models. The ML models' efficiency will be evaluated to achieve a
final one. Rasults: An initial data preprocessing was performed using the Python module known as
Pandas Profiling with data from the first evaluation of 38 diabetic patients, with a mean age of
78.5 years (± 6.64), with 29 (76.3 %) women. Correlation trends are observed between the nominal
variables "exercise timeR (min) and "taking medication correctly" as well as "list, dosage and
medications" and "diabetes contraiR. The correlation between ali variables shown positive and
negative correlations between gait variables. There is also a correlation tendency between gait
variables and anthropometric variables, such as weight, height, and body fat. Likewise, a positive
correlation trend between balance variables and glycemic indexes.

BANKING MEMBERS:
Presidente - 2374822 - FABRICIA AZEVEDO DA COSTA CAVALCANTI
Externo ao Programa - 1763991 - RUMMENIGGE RUDSON DANTAS
Interna - 2319151 - TATIANA SOUZA RIBEIRO