Poster
in
Workshop: Workshop on Learning from Time Series for Health
Temporally Multi-Scale Sparse Self-Attention for Physical Activity Data Imputation
Hui Wei · Maxwell Xu · Colin Samplawski · James Rehg · Santosh Kumar · Benjamin M Marlin
Keywords: [ autocorrelation ] [ time series ] [ missing data ] [ step count ] [ sparsity ] [ self-attention ]
Wearable sensors enable health researchers to continuously collect data pertaining to the physiological state of individuals in real-world settings. However, such data can be subject to extensive missingness due to a complex combination of factors. In this work, we study the problem of imputation of missing step count data, one of the most ubiquitous forms of wearable sensor data. We construct a novel and large scale data set consisting of a training set with over 3 million hourly step count observations and a test set with over 2.5 million hourly step count observations. We propose an autocorrelation-informed sparse self-attention model for this task that captures the temporally multi-scale nature of step count data. We assess the performance of the proposed model relative to baselines based on different missing rates.