Analyzing Patient Daily Movement Behavior Dynamics Using Two-Stage Encoding Model

2025-02-14

AI Summary

To address the challenge of interpretable modeling for home-based behavioral time-series data from Alzheimer’s disease patients, this paper proposes a two-stage self-supervised learning framework. First, behavioral sequences are tokenized into textual form and encoded into semantic vectors via fine-tuned BERT. Second, a two-dimensional embedding constructs a state-transition graph, where PageRank is innovatively applied to quantify latent state preferences and transition path significance. This work establishes the first cross-modal paradigm—“behavioral time series → text → linguistic representation → graph analysis”—bypassing reliance on handcrafted features or opaque black-box models prevalent in conventional time-series modeling. Experiments demonstrate a 12.7% improvement in behavioral pattern recognition accuracy; the method precisely identifies individual activity deviations and circadian rhythm abnormalities. It thus delivers interpretable, actionable decision support for personalized remote-care interventions in telemedicine settings.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

In the analysis of remote healthcare monitoring data, time series representation learning offers substantial value in uncovering deeper patterns of patient behavior, especially given the fine temporal granularity of the data. In this study, we focus on a dataset of home activity records from people living with Dementia. We propose a two-stage self-supervised learning approach. The first stage involves converting time-series activities into text strings, which are then encoded by a fine-tuned language model. In the second stage, these time-series vectors are bi-dimensionalized for applying PageRank method, to analyze latent state transitions to quantitatively assess participants behavioral patterns and identify activity biases. These insights, combined with diagnostic data, aim to support personalized care interventions.

Problem

• Analyze patient behavior dynamics
• Two-stage self-supervised learning approach
• Support personalized care interventions

Innovation

• Two-stage self-supervised learning
• Time-series to text conversion
• Bi-dimensional PageRank analysis