The proposed framework, EGCN, integrates temporal entropy analysis, graph-based modeling, clustering, and forecasting to detect anomalies and predict future trends in spatiotemporal data.

The proposed framework, EGCN, integrates temporal entropy analysis, graph-based modeling, clustering, and forecasting to detect anomalies and predict future trends in spatiotemporal data.

<p>It begins with multi-dimensional time-series data for <i>m</i> entities, each with features such as price (<i>p</i>), volume (<i>v</i>) and geospatial data (<i>s</i>), sampled at <i>n</i> time points (). Entropy values are computed...

Full description

Saved in:

Bibliographic Details
Main Author:	Dat Le (10650470) (author)
Other Authors:	Sutharshan Rajasegarar (14481165) (author), Wei Luo (80175) (author), Thanh Thi Nguyen (13110078) (author), Nhi Vo (19757889) (author), Quang Nguyen (565637) (author), Maia Angelova (4285048) (author)
Published:	2025
Subjects:	Medicine Cell Biology Pharmacology Developmental Biology Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified indicating superior sensitivity anomalous region detection anomalous pattern detection anomalous &# 8212 48 months ), time series forecasting tested forecasting models specific forecasting framework reduced prediction stability improved forecasting accuracy enhanced market insights applies forecasting models 48 months ). us ), 14 12 months ), egcn remains stable clustering anomalies separately anomalous regions independently australia ), highlighting us ), 12 market deviations forecasting horizons forecasting errors australia ). uk ), treat normal term forecasts results highlight reduces mse reduced across propose egcn policy changes novel cluster normal regions lowest errors lowest error lowering mse inherently dynamic higher errors first detects extended periods expected trends error accumulation egcn achieves economic fluctuations demographic shifts decreasing mase
Tags:	Add Tag No Tags, Be the first to tag this record!

Be the first to leave a comment!

Cannot write session to /tmp/vufind_sessions/sess_ukcveeqrpfhvpg3p9gg016u9g4