Comparative Study of Pressure Ulcer classification based on Braden scale in the Community-Dwelling Older Adults
Desri Kristina (a*), Husneni Mukhtar (b), Sheizi Prista Sari (c), Eka Afrima Sari (d)
a, b) Fakultas Teknik Elektro Telkom University, Bandung, Indonesia
c, d) Fakultas Ilmu Keperawatan Universitas Padjadjaran, Bandung, Indonesia
Abstract
A problem that is often found in the elderly is pressure ulcers cause decreased quality of life elderly. One way to recognize the risk of pressure ulcer by Braden Scale analysis. Braden Scale has been tested with the level of reliability and validity with various types of hospitals and patients. The purpose of this research is to compare several classification methods to classify pressure ulcers in dataset. The classification methods used are Logistic Regression, Quadratic Discriminant Analysis, K Nearest Neighbor, Support Vector Machine, Naive Bayes, Gaussian Process Classifier and Multi-Layer Perceptron Classifier. The method used to evaluate performance or algorithm is k-fold cross validation with k=10.Based on the analysis results obtained that Quadratic Discriminant Analysis produces the highest accuracy, with an accuracy rate is 94.12%.
Keywords: Classification; Machine learning; Pressure ulcer, Quadratic discriminant analysis
Topic: Computers, Internet, Multimedia in Engineering Education