When to Update Sequential Patterns in Stream Data?

Title: When to Update Sequential Patterns in Stream Data?

Abstract: This research investigates the optimal time to update sequential patterns in stream data. It proposes a method to determine when to update sequential patterns by balancing the performance of increasingly updating algorithms and the difference between old and new patterns. Experiments on two data sets show that the optimal size of incremental windows to update is approximately 15-25% of the initial window size.

Main Research Question: How can we determine the optimal time to update sequential patterns in stream data while maintaining the performance of increasingly updating algorithms?

Methodology: The study first defines a distance measure to determine the difference between old and new sequential patterns. It then proposes a method to decide when to update sequential patterns by considering the trade-off between the performance of increasingly updating algorithms and the difference in patterns.

Results: Experiments on two data sets demonstrate that as the size of incremental windows grows, the values of the speedup and the differences in patterns decrease and increase, respectively. The study shows that the optimal size of incremental windows to update is around 15-25% of the initial window size.

Implications: This research provides a practical solution to the problem of determining when to update sequential patterns in stream data. It helps data miners balance the need for updating patterns with the computational resources available. The proposed method can be applied to various domains, such as finance, healthcare, and e-commerce, where real-time data analysis is crucial.

Link to Article: https://arxiv.org/abs/0203028v1 Authors: arXiv ID: 0203028v1