Publication: Prediction of standing height for hospitalized elderly using multilayer perceptron network
Loading...
Date
2010-04-01
Authors
Nik Sulaiman, Nik Suraya
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Artificial intelligence (AI) method such as artificial neural network (ANN) is widely been applied in a huge number of various applications, especially in medical approaches. The measurement of height for hospitalized and bedridden patients is very needed as it is important for clinical purposes such as for calculating body mass index (BMI) and evaluating the nutritional status. But, it becomes problem for those who cannot stand straight. So, demi-span measurement becomes an alternative way to measures stature in elderly people. This project aims to predict standing height using multilayer perceptron (MLP) network. The data set used for this project was collected from 360 elderly patients, with 180 patients are males and another 180 patients are females. Levenberg-Marquardt (LM) and Bayesian Regularization Back propagation (BR) algorithms are used in this project. The result or performance of both networks was compared according to accuracy and mean square error (MSE). The number of hidden neuron used was also investigated. The results suggested that BR has better performance as it could achieve higher accuracy than LM algorithm and it has low MSE value.