Open AccessFall Detection for Elderly Person using Neuro-fuzzy System and Wavelet Transformation
Author(s) -
Sang-Hong Lee*
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.l3200.1081219
Subject(s) - test set , backpropagation , set (abstract data type) , artificial intelligence , wavelet , computer science , artificial neural network , transformation (genetics) , pattern recognition (psychology) , carry (investment) , training set , fuzzy set , wavelet transform , machine learning , fuzzy logic , biochemistry , chemistry , finance , economics , gene , programming language
This study proposes a new methodology to detect falls and non-falls using a Neural Network with Weighted Fuzzy Membership Functions (NEWFM). Dataset acquired from subjects was applied to NEWFM after carrying out wavelet transforms. In order to test the performance evaluation of the fall detection by the NEWFM, the dataset was separated test set and training set at 2 to 8 and 5 to 5 ratios to carry out experiments. Based on the performance evaluation of the NEWFM, the sensitivity, accuracy, and specificity were shown to be 94.67%, 91.86% and 89.41%, respectively when the test set to the training set at the ratio was 2 to 8 and 91%, 91% and 91%, respectively, when the test set to the training set at the ratio was 5 to 5. This study also compares the performance evaluation of backpropagation (BP) and that of NEWFM.