Title
Self-Reported Well-Being Score Modelling And Prediction: Proof-Of-Concept Of An Approach Based On Linear Dynamic Systems
Abstract
Assessment and recognition of perceived wellbeing has wide applications in the development of assistive healthcare systems for people with physical and mental disorders. In practical data collection, these systems need to be less intrusive, and respect users' autonomy and willingness as much as possible. As a result, self-reported data are not necessarily available at all times. Conventional classifiers, which usually require feature vectors of a prefixed dimension, are not well suited for this problem. To address the issue of non-uniformly sampled measurements, in this study we propose a method for the modelling and prediction of self-reported well-being scores based on a linear dynamic system. Within the model, we formulate different features as observations, making predictions even in the presence of inconsistent and irregular data. We evaluate the proposed method with synthetic data, as well as real data from two patients diagnosed with cancer. In the latter, self-reported scores from three well-being-related scales were collected over a period of approximately 60 days. Prompted each day, the patients had the choice whether to respond or not. Results show that the proposed model is able to track and predict the patients' perceived well-being dynamics despite the irregularly sampled data.
Year
DOI
Venue
2017
10.1109/EMBC.2017.8037292
2017 39TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC)
Field
DocType
Volume
Data collection,Signal processing,Data mining,Feature vector,Computer science,Autonomy,Proof of concept,Synthetic data,Artificial intelligence,Health informatics,Machine learning,Dynamical system
Conference
2017
ISSN
Citations 
PageRank 
1094-687X
0
0.34
References 
Authors
2
5
Name
Order
Citations
PageRank
Xinyang Li100.34
Riccardo Poli22589308.79
Gaetano Valenza331448.21
Enzo Pasquale Scilingo439498.07
luca citi516827.88