The Effect of Personalization on Smartphone-Based Fall Detectors.

Carlos Medrano,Raúl Igual,Ángel Sánchez,Manuel Castro,Inmaculada Plaza

doi:10.3390/s16010117

Abstract

The risk of falling is high among different groups of people, such as older people, individuals with Parkinson's disease or patients in neuro-rehabilitation units. Developing robust fall detectors is important for acting promptly in case of a fall. Therefore, in this study we propose to personalize smartphone-based detectors to boost their performance as compared to a non-personalized system. Four algorithms were investigated using a public dataset: three novelty detection algorithms—Nearest Neighbor (NN), Local Outlier Factor (LOF) and One-Class Support Vector Machine (OneClass-SVM)—and a traditional supervised algorithm, Support Vector Machine (SVM). The effect of personalization was studied for each subject by considering two different training conditions: data coming only from that subject or data coming from the remaining subjects. The area under the receiver operating characteristic curve (AUC) was selected as the primary figure of merit. The results show that there is a general trend towards the increase in performance by personalizing the detector, but the effect depends on the individual being considered. A personalized NN can reach the performance of a non-personalized SVM (average AUC of 0.9861 and 0.9795, respectively), which is remarkable since NN only uses activities of daily living for training.

Highlights

Falls are an important public health concern, especially for some groups of people
A personalized Nearest Neighbor (NN) can reach the performance of a non-personalized Support Vector Machine (SVM), which is remarkable since NN only uses activities of daily living for training
Regarding the usability problems that may arise, fall detection applications can be programmed in such a way that they can override the normal operation of the elements of the smartphone that can represent a usability barrier for low-skilled users [18]

Summary

Introduction

Falls are an important public health concern, especially for some groups of people. It is estimated that about one-third of those over 65 years old experience one or more falls every year, but this fraction is even higher among the oldest ones [1]. Regarding the usability problems that may arise, fall detection applications can be programmed in such a way that they can override the normal operation of the elements of the smartphone that can represent a usability barrier for low-skilled users [18] They can boost the factors that facilitate the adoption of technology by older people [5], such as usability, control, feedback and cost. ADL can be recorded in real life, as opposed to in laboratory experimental falls They are good candidates for personalization, since they can be re-trained with the new data of a given user. Personalization using real-world falls from a given person is unrealistic, but at least SVM can be adapted with new, true ADL from a given user recorded while he/she carries the phone

Dataset

Algorithms and Their Evaluation

D P Nk pCq

Comparison between Novelty Detectors

Personalization

Discussion