Abstract
Sensor-based human activity recognition (HAR) has become a popular research topic because of its wide applications. Conventional machine learning approaches have made enormous progress in the past years. However, those methods rely on handcrafted features that are incapable of handling complex activities, especially with high dimensional sensor data. Deep learning technology, together with its various models, is one of the most accurate methods of working on activity data. In this paper, we propose an attention-based Long Short Term Memory (LSTM) network for wearable human activity recognition. Specifically, we construct an LSTM network to model the sensor readings, which has been proved to be very effective for time sequences. Then, we introduce the attention mechanism for the base LSTM network to learn which parts of the raw sensor data are more important for determining the overall activities. When tested with the Opportunity data set, the F1-score is increased by 2.6%, compared with baseline LSTM results.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
