.png){kind=icon}
.png){kind=icon}
Return
Time-dependent frequent sequence mining-based survival analysis
.png){kind=icon}
OA
AI
DOI：10.1016/j.knosys.2024.111885.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Frequent sequence mining is a valuable technique for identifying patterns and co-occurrences in event sequences. However, traditional approaches often neglect the temporal information associated with events, limiting their ability to capture the dynamics of event sequences. In this study, we propose a methodology that integrates frequent sequence mining with survival analysis to address this limitation. Frequent sequence mining captures the order and frequency of occurrence of typical events, while association rules highlight the relevant ones. In addition, survival analysis provides comprehensive temporal information between them. The approach also handles competing risks simultaneously, ensuring unbiased results. The output of the method is sequences of distribution functions of the elapsed time between the frequent and relevant events, which describe the time-varying confidence of the frequent sequences. The method also presents how time-varying confidence functions can be enhanced by explanatory variables and how their confidence interval can be determined using the bootstrapping method. The applicability of the approach is demonstrated using clinical data, specifically focusing on disease sequences.
Keywords:
Frequent sequence mining
Survival analysis
Competing risks
Time-varying confidence function
Bootstrapping
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
K
IF:
7.6
Papers:
1.2W
Citations:
4.5W