How to estimate the emotions hidden behind spatio-temporal
Keywords: Behavior pattern analysis, Emotional perception, Spatiotemporal trajectory
Abstract: Emotion estimation of online spatiotemporal behavior is a
technique for studying mental health and its changing laws
based on spatiotemporal trajectory data of objects. According
to WHO data, the proportion of patients with depression
worldwide is as high as 3.7%, and mental health detection
technology has become a new hotspot in current international
research. Traditional technologies mainly collect physiological data such as heart rate, blood pressure, blood oxygen and
sleep through wearable devices (such as wristbands) to
achieve online analysis of mental health levels. However, the
low measurement accuracy of wearable devices makes it difficult to meet the quality requirements for emotion estimation.
More importantly, emotional changes are not only affected
by physiological factors, but social factors are more important. This paper studies the relationship between the object's spatiotemporal behavior and emotional state, focusing
on the mechanism of the object's social behavior pattern and
its changes on emotional changes. A social activity pattern
extraction method based on spatio-temporal trajectory data is
proposed, a social activity sequence expression model of the
subject's daily behavior is established, and the mapping relationship between the social activity sequence and the emotional index under multi-resolution is explored. The experimental results show that the object's social and social activity
patterns are closely related to its emotional index. The proposed SADS emotion estimation model is better than the
baseline paper on both SAPD22111510 and SAPD23031530
datasets, with an average increase in accuracy of 3.9% and
8.1% respectively. For the first time, the paper expands the
research object of online emotion estimation from traditional
spatiotemporal behavior to social behavior pattern research,
which provides new research ideas and technical approaches
for online emotion estimation research.
Primary Area: societal considerations including fairness, safety, privacy
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 1231
Loading