Temporal Information Enhanced Association Rule Discovery and Evolution Pattern Identification in Complex Pharmaceutical Transaction Data
Journal of Disease and Public Health · 2026
DOI: 10.71052/jdph/fhin5235 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Disease and Public Health, 2026. Verfasst von Haiyang Chen. Die Arbeit nennt 11 Literaturangaben.
Abstract (Originalsprache)
Objective
To bring temporal information into association rule mining on complex pharmaceutical transaction data, so that co-prescribing rules and their evolution over time can both be identified.
Methods
Transaction records are reorganized into time-stamped drug-set sequences at a selected Anatomical Therapeutic Chemical (ATC) abstraction level. Temporal association rules with interval constraints are extracted using a residual-leverage measure that shrinks toward the null hypothesis. Each rule is represented as a trajectory of support, confidence and leverage across sliding windows and clustered into five evolution modes: stable, emerging, decaying, periodic and abrupt.
Results
On a provincial procurement and insurance settlement dataset, the method compresses the rule space while preserving coverage. It also exposes emerging polypharmacy combinations earlier than cross-sectional screening.
Conclusion
It provides an interpretable instrument for rational drug use surveillance, chronic disease monitoring, drug interaction signal detection and insurance fund risk warning.
Zitieren
APA
Chen, H. (2026). Temporal Information Enhanced Association Rule Discovery and Evolution Pattern Identification in Complex Pharmaceutical Transaction Data. Journal of Disease and Public Health. https://doi.org/10.71052/jdph/fhin5235
Vancouver
Chen H. Temporal Information Enhanced Association Rule Discovery and Evolution Pattern Identification in Complex Pharmaceutical Transaction Data. Journal of Disease and Public Health. 2026. doi:10.71052/jdph/fhin5235
BibTeX
@article{Chen2026,
author = {Chen, Haiyang},
title = {Temporal Information Enhanced Association Rule Discovery and Evolution Pattern Identification in Complex Pharmaceutical Transaction Data},
journal = {Journal of Disease and Public Health},
year = {2026},
doi = {10.71052/jdph/fhin5235}
}
EUVIMED