Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

SummerTime: Variable-length Time Series Summarization with Application to Physical Activity Analysis

SummerTime: Variable-length Time Series Summarization with Application to Physical Activity Analysis SummerTime seeks to summarize global time-series signals and provides a fixed-length, robust representation of the variable-length time series. Many machine learning methods depend on data instances with a fixed number of features. As a result, those methods cannot be directly applied to variable-length time series data. Existing methods such as sliding windows can lose minority local information. Summarization conducted by the SummerTime method will be a fixed-length feature vector which can be used in place of the time series dataset for use with classical machine learning methods. We use Gaussian Mixture models (GMM) over small same-length disjoint windows in the time series to group local data into clusters. The time series’ rate of membership for each cluster will be a feature in the summarization. By making use of variational methods, GMM converges to a more robust mixture, meaning the clusters are more resistant to noise and overfitting. Further, the model is naturally capable of converging to an appropriate cluster count. We validate our method on a challenging real-world dataset, an imbalanced physical activity dataset with a variable-length time series structure. We compare our results to state-of-the-art studies and show high-quality improvement by classifying with only the summarization. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ACM Transactions on Computing for Healthcare (HEALTH) Association for Computing Machinery

SummerTime: Variable-length Time Series Summarization with Application to Physical Activity Analysis

Download PDF
15 pages

Loading next page...
 
/lp/association-for-computing-machinery/summertime-variable-length-time-series-summarization-with-application-Q2roFKR4si
Publisher
Association for Computing Machinery
Copyright
Copyright © 2022 Association for Computing Machinery.
ISSN
2691-1957
eISSN
2637-8051
DOI
10.1145/3532628
Publisher site
See Article on Publisher Site

Abstract

SummerTime seeks to summarize global time-series signals and provides a fixed-length, robust representation of the variable-length time series. Many machine learning methods depend on data instances with a fixed number of features. As a result, those methods cannot be directly applied to variable-length time series data. Existing methods such as sliding windows can lose minority local information. Summarization conducted by the SummerTime method will be a fixed-length feature vector which can be used in place of the time series dataset for use with classical machine learning methods. We use Gaussian Mixture models (GMM) over small same-length disjoint windows in the time series to group local data into clusters. The time series’ rate of membership for each cluster will be a feature in the summarization. By making use of variational methods, GMM converges to a more robust mixture, meaning the clusters are more resistant to noise and overfitting. Further, the model is naturally capable of converging to an appropriate cluster count. We validate our method on a challenging real-world dataset, an imbalanced physical activity dataset with a variable-length time series structure. We compare our results to state-of-the-art studies and show high-quality improvement by classifying with only the summarization.

Journal

ACM Transactions on Computing for Healthcare (HEALTH) – Association for Computing Machinery

Published: Dec 7, 2022

Keywords: Time series

References

  • Automatic identification of physical activity types and sedentary behaviors from triaxial accelerometer: Laboratory-based calibrations are not enough
    [,
  • Time series representation and similarity based on local autopatterns
    [,
  • Mixture models and em
    [,
  • Neural networks
    [,
  • A deep learning architecture for temporal sleep stage classification using multivariate and multimodal time series
    [,
  • A novel method for using accelerometer data to predict energy expenditure
    [,
  • Evaluation of artificial neural network algorithms for predicting mets and activity type from accelerometer data: validation on an independent sample
    [,
  • Toeplitz inverse covariance-based clustering of multivariate time series data
    [,
  • Classification of time series by shapelet transformation
    [,
  • Multivariate LSTM-FCNS for time series classification
    [,
  • R&P: An low-cost device-free activity recognition for E-health
    [,
  • version R2018a
    [,
  • Bipart: Learning block structure for activity detection
    [,
  • Variational learning for Gaussian mixture models
    [,
  • Two-stage topic extraction model for bibliometric data analysis based on word embeddings and clustering
    [,
  • Sentiment analysis on product reviews based on weighted word embeddings and deep neural networks
    [,
  • Sentiment analysis on massive open online course evaluations: A text mining and deep learning approach
    [,
  • Weighted word embeddings and clustering-based identification of question topics in MOOC discussion forum posts
    [,
  • A term weighted neural language model and stacked bidirectional LSTM based framework for sarcasm identification
    [,
  • Mining opinions from instructor evaluation reviews: A deep learning approach
    [,
  • Sentiment analysis in Turkish based on weighted word embeddings
    [,
  • Topic-enriched word embeddings for sarcasm identification
    [,
  • A multiobjective weighted voting ensemble classifier based on differential evolution algorithm for text sentiment classification
    [,
  • Ensemble of keyword extraction methods and classifiers in text classification
    [,
  • An improved ant algorithm with LDA-based representation for text document clustering
    [,
  • Dynamic time warping averaging of time series allows faster and more accurate classification
    [,
  • An artificial neural network to estimate physical activity energy expenditure and identify physical activity type from an accelerometer
    [,
  • Gait-based identification for elderly users in wearable healthcare systems
    [,
  • Artificial neural networks to predict activity type and energy expenditure in youth
    [,
  • Time series shapelets: A new primitive for data mining
    [,
  • Motionword: An activity recognition algorithm based on intelligent terminal and cloud
    [,
  • Ultigesture: A wristband-based platform for continuous gesture control in healthcare
    [,
  • Ultigesture: A wristband-based platform for continuous gesture control in healthcare
    Zhao, Hongyang; Wang, Shuangquan; Zhou, Gang; Zhang, Daqing