Selecting Orientation-Insensitive Features for Activity Recognition from Accelerometers

Abstract

Activity recognition from sensors is a classification problem over time-series data. Some research in the area utilize time and frequency domain handcrafted features that differ between datasets. Another categorically different approach is to use deep learning methods for feature learning. This paper explores a middle ground in which an off-theshelf feature extractor is used to generate a large number of candidate timedomain features followed by a feature selector that was designed to reduce the bias toward specific classification techniques. Moreover, this paper advocates the use of features that are mostly insensitive to sensor orientation and show their applicability to the activity recognition problem. The proposed approach is evaluated using six different publicly available datasets collected under various conditions using different experimental protocols and shows comparable or higher accuracy than state-of-the-art methods on most datasets but usually using an order of magnitude fewer features.

Predicting contact-without-connection defects on printed circuit boards employing ball grid array package types: a data analytics case study in the smart manufacturing environment

“…It is not a tool for time series regression but rather to help utilize time series for supervised machine learning problems in an exploratory fashion. There is evidence to suggest that specially-crafted features tailored to the problem of interest will outperform the standard set of features calculated by TSFRESH [33]. However, there do not exist any specially-crafted features for this problem.…”

Significant Features for Human Activity Recognition Using Tri-Axial Accelerometers

“…Currently, the accelerometer and gyroscope are the most widely used wearable sensors for PA monitoring. They have been used in a variety of applications, including heath monitoring, fall detection [ 7 ], human movement monitoring [ 8 ], gait analysis [ 9 ], disease progression assessment [ 10 ], and activity recognition [ 11 ]. Accelerometers are a popular choice of sensor due to their practicality, low cost and reliability, and have been employed in many systems to classify Activities of Daily Living (ADL), also known as Human Activity (HA) [ 5 ].…”