Abstract
In this paper, we present and benchmark FilterNet, a flexible deep learning architecture for time series classification tasks, such as activity recognition via multichannel sensor data. It adapts popular convolutional neural network (CNN) and long short-term memory (LSTM) motifs which have excelled in activity recognition benchmarks, implementing them in a many-to-many architecture to markedly improve frame-by-frame accuracy, event segmentation accuracy, model size, and computational efficiency. We propose several model variants, evaluate them alongside other published models using the Opportunity benchmark dataset, demonstrate the effect of model ensembling and of altering key parameters, and quantify the quality of the models’ segmentation of discrete events. We also offer recommendations for use and suggest potential model extensions. FilterNet advances the state of the art in all measured accuracy and speed metrics when applied to the benchmarked dataset, and it can be extensively customized for other applications.
Highlights
While the time series classification algorithm described in this work is applicable to many of these domains, it was developed as an activity recognition algorithm, and we present and benchmark it here primarily in that context
We have found that 3exhibited the highest accuracy, while smaller and simpler variants were generally faster but less to-5-fold Multi-Scale CNN/LSTM (ms-C/L) ensembles perform well on many tasks and datasets, especially if inference speed accurate
long short-term memory (LSTM) implementation as we have done in this work, and by using an architecture with efficiency drawbacks ofstride executing on GPUs can be substantially by using a GPUa large output-to-input ratio LSTMs so that the input sequence to the LSTMmitigated layer is shorter
Summary
Time series classification is a challenging problem in numerous fields [1], including finance [2], cyber security [3], electronic health record analysis [4], acoustic scene classification [5], and electroencephalogram (EEG)-based brain computer interfaces [6], and it is a central challenge in the field of activity recognition [7].Numerous time series classification algorithms have been proposed [8,9], and the diversity of time series classification problems is evident in dataset repositories, such as the UCR Time Series Archive [10] or the UCI Machine Learning Repository [11]. Time series classification is a challenging problem in numerous fields [1], including finance [2], cyber security [3], electronic health record analysis [4], acoustic scene classification [5], and electroencephalogram (EEG)-based brain computer interfaces [6], and it is a central challenge in the field of activity recognition [7]. While the time series classification algorithm described in this work is applicable to many of these domains, it was developed as an activity recognition algorithm, and we present and benchmark it here primarily in that context. AR is used in applications as diverse as smart homes [12,13], gesture recognition [14], computer control interfaces [15], health monitoring [16], and home behavior analysis [17]. Other sensors may include microphones, pressure sensors, and various object-mounted sensors that detect object displacement
Sign-in/Register to view highlights & summary 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.
