Performance Analysis of Boosting Classifiers in Recognizing Activities of Daily Living.

Saifur Rahman,Shumayla Yaqoob,Mohsin Raza,Muhammad Awais,Muhammad Irfan,Khawaja Moyeezullah Ghori

doi:10.3390/ijerph17031082

Abstract

Physical activity is essential for physical and mental health, and its absence is highly associated with severe health conditions and disorders. Therefore, tracking activities of daily living can help promote quality of life. Wearable sensors in this regard can provide a reliable and economical means of tracking such activities, and such sensors are readily available in smartphones and watches. This study is the first of its kind to develop a wearable sensor-based physical activity classification system using a special class of supervised machine learning approaches called boosting algorithms. The study presents the performance analysis of several boosting algorithms (extreme gradient boosting—XGB, light gradient boosting machine—LGBM, gradient boosting—GB, cat boosting—CB and AdaBoost) in a fair and unbiased performance way using uniform dataset, feature set, feature selection method, performance metric and cross-validation techniques. The study utilizes the Smartphone-based dataset of thirty individuals. The results showed that the proposed method could accurately classify the activities of daily living with very high performance (above 90%). These findings suggest the strength of the proposed system in classifying activity of daily living using only the smartphone sensor’s data and can assist in reducing the physical inactivity patterns to promote a healthier lifestyle and wellbeing.

Highlights

World health organizations defined physical activity as any body movement that requires energy expenditure to perform any task originated through the musculoskeletal system [1]
This study investigates the performance of recently developed Catboost classifiers for physical activity classification (PAC)
F-measure metric. by Theusing performance of each was computed using both feature sets, without feature selection, all the feature set classifier obtained originally (All feat), and with feature selection, feature selection, by using all the feature set obtained originally (All feat), and with feature selection, by only utilizing the uncorrelated feature set obtained through Correlation-based features (CFS) approach

Summary

Introduction

World health organizations defined physical activity as any body movement that requires energy expenditure to perform any task originated through the musculoskeletal system [1]. Physical activity is quite essential for human beings to carry on their daily living routine work. Activities are the movements that the body does all day long. Picking up fruits, cleaning the house, sitting, standing, walking and lying are examples of activities of daily living (ADLs). Physical activity should not be confused with exercise since it is a sub-branch of physical activity. Exercise is a set of well-planned, structured and repetitive actions where we plan to move a specific part of the body, such as lower or upper limbs, and the movement is repeated, for example, lifting the weight ten times in a row

Objectives

Methods

Findings

Discussion

Conclusion