NUMERO: #20240203

Título: ACTINIC CHEILITIS MONITORING SYSTEM BASED ON RANDOM FOREST: A MACHINE LEARNING APPROACH

Nome do Apresentador: Ivan José CORREIA-NETO

Categoria do Trabalho: Painel de pesquisa científica (PPC)

Área Temática: Estomatologia

Resumo: Objective: This study investigated the potential of a random forest (RF) algorithm for the clinical diagnosis and evolution of actinic cheilitis (AC). Methods: The dataset consists of 237 patients, with 7 cases which progressed to spindle cell carcinoma. To balance the data, a computational technique was applied to virtually increase the number of positive instances, based on the range of values of existing parameters. Subsequently, a RF model with four trees and ten estimators, was trained and tested using 5-fold cross-validation. Results: A balanced 96.08% accuracy, 98.67% sensitivity, 92.06% specificity, 92.31% precision, 96% f1-score, and 96.43% AUC were obtained. Conclusion: RF is a promising tool for assisting in the clinical diagnosis and evolution of AC, offering valuable insights for controlling the non-malignant transformation of AC, however an increase in the database is required to bring more class variability information; additionally, other models should be considered for training and validation.

Autor 1:  Ivan José CORREIA-NETO

E-mail 1:  [email protected]

Autor 2:   Alex Franco da COSTA

E-mail 2:  [email protected]

Autor 3 :  Anna Luíza Damaceno ARAÚJO

E-mail 3:  [email protected]

Autor 4:  Thiago Martini PEREIRA

E-mail 4:  [email protected]

Autor 5:  Alan Roger SANTOS-SILVA

E-mail 5:  [email protected]

Autor 6:  Matheus Cardoso MORAES

E-mail 6:  [email protected]

Autor 7:  Márcio Ajudarte LOPES

E-mail 7:  [email protected]