Classification of 24-hour movement behaviors from wrist-worn accelerometer data: from handcrafted features to deep learning techniques

2025-09-10

AI Summary

This study addresses 24-hour physical behavior recognition using wrist-worn accelerometer data, targeting fine-grained classification of sleep, sedentary behavior, light-intensity physical activity (LPA), and moderate-to-vigorous physical activity (MVPA). Raw acceleration signals were segmented via sliding windows and fed into both deep learning models (LSTM, BiLSTM, GRU, 1D-CNN) and traditional machine learning classifiers (Random Forest, SVM, XGBoost), evaluated on both raw time-series inputs and handcrafted 104-dimensional features. Results demonstrate, for the first time empirically, that end-to-end deep models operating directly on raw signals achieve a mean accuracy of 85%, significantly outperforming traditional approaches on engineered features (70–81%). However, substantial confusion persists between LPA and MVPA classes. The work elucidates both the advantages—superior discriminative capacity through automatic feature learning—and limitations—persistent ambiguity in distinguishing activity intensities—of end-to-end modeling for complex behavioral classification. It provides methodological guidance and empirical validation for wearable-driven, high-fidelity physical behavior assessment.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Purpose: We compared the performance of deep learning (DL) and classical machine learning (ML) algorithms for the classification of 24-hour movement behavior into sleep, sedentary, light intensity physical activity (LPA), and moderate-to-vigorous intensity physical activity (MVPA). Methods: Open-access data from 151 adults wearing a wrist-worn accelerometer (Axivity-AX3) was used. Participants were randomly divided into training, validation, and test sets (121, 15, and 15 participants each). Raw acceleration signals were segmented into non-overlapping 10-second windows, and then a total of 104 handcrafted features were extracted. Four DL algorithms-Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM), Gated Recurrent Units (GRU), and One-Dimensional Convolutional Neural Network (1D-CNN)-were trained using raw acceleration signals and with handcrafted features extracted from these signals to predict 24-hour movement behavior categories. The handcrafted features were also used to train classical ML algorithms, namely Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Logistic Regression (LR), Artificial Neural Network (ANN), and Decision Tree (DT) for classifying 24-hour movement behavior intensities. Results: LSTM, BiLSTM, and GRU showed an overall accuracy of approximately 85% when trained with raw acceleration signals, and 1D-CNN an overall accuracy of approximately 80%. When trained on handcrafted features, the overall accuracy for both DL and classical ML algorithms ranged from 70% to 81%. Overall, there was a higher confusion in classification of MVPA and LPA, compared to sleep and sedentary categories. Conclusion: DL methods with raw acceleration signals had only slightly better performance in predicting 24-hour movement behavior intensities, compared to when DL and classical ML were trained with handcrafted features.

Problem

• Classifying 24-hour movement behaviors using accelerometer data
• Comparing deep learning and classical machine learning algorithms
• Evaluating performance for sleep, sedentary, LPA, and MVPA categories

Innovation

• Deep learning algorithms using raw acceleration signals
• Comparison with classical ML on handcrafted features
• LSTM/BiLSTM/GRU achieving 85% classification accuracy