View in EDS

Rule-based interpretable sequence clustering.

Saved in:
Bibliographic Details
Title: Rule-based interpretable sequence clustering.
Authors: Liu, Yushuang1 (AUTHOR) lys2665@163.com, Li, Pengju1 (AUTHOR) pj_li@qq.com, Dong, Junjie1,2 (AUTHOR) jd445@qq.com, Jiang, Mudi1 (AUTHOR) 792145962@qq.com, Liu, Xinying1 (AUTHOR) 72317011@mail.dlut.edu.cn, Hu, Lianyu1,3 (AUTHOR) hly4ml@gmail.com, He, Zengyou1 (AUTHOR) zyhe@dlut.edu.cn
Source: Knowledge & Information Systems. 6/27/2026, Vol. 68 Issue 1, p1-30. 30p.
Subjects: Rule-based programming, Clustering algorithms, Association rule mining, Optimization algorithms, Machine learning, Data mining, Decision trees
Abstract: During the past decades, many effective algorithms for clustering discrete sequences have be presented. However, most of these existing clustering methods lack interpretability, i.e., the capability of explaining the identified clusters in an intuitive manner. To our knowledge, there is only one interpretable sequence clustering method in the literature so far, which employs a pattern-based decision tree for explaining sequence clusters. Such a tree-based model may become less interpretable with the increase in cluster number and each cluster is characterized by a conjunction of sequential patterns. To address this limitation, we propose a rule-based interpretable sequence clustering algorithm, consisting of two main components: discriminative rule mining and rule set optimization. In the rule set, each rule is associated with only one pattern, ensuring a high level of interpretability. Experiments conducted on 13 real-world datasets demonstrate that our method achieves competitive accuracy compared with state-of-the-art interpretable and non-interpretable sequence clustering algorithms. At the same time, it exhibits significant advantages in interpretability: each rule is non-conjunctive, the rule set is compact, the average pattern length is short, and the resulting explanations are more concise and intuitive. [ABSTRACT FROM AUTHOR]
Copyright of Knowledge & Information Systems is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:During the past decades, many effective algorithms for clustering discrete sequences have be presented. However, most of these existing clustering methods lack interpretability, i.e., the capability of explaining the identified clusters in an intuitive manner. To our knowledge, there is only one interpretable sequence clustering method in the literature so far, which employs a pattern-based decision tree for explaining sequence clusters. Such a tree-based model may become less interpretable with the increase in cluster number and each cluster is characterized by a conjunction of sequential patterns. To address this limitation, we propose a rule-based interpretable sequence clustering algorithm, consisting of two main components: discriminative rule mining and rule set optimization. In the rule set, each rule is associated with only one pattern, ensuring a high level of interpretability. Experiments conducted on 13 real-world datasets demonstrate that our method achieves competitive accuracy compared with state-of-the-art interpretable and non-interpretable sequence clustering algorithms. At the same time, it exhibits significant advantages in interpretability: each rule is non-conjunctive, the rule set is compact, the average pattern length is short, and the resulting explanations are more concise and intuitive. [ABSTRACT FROM AUTHOR]
ISSN:02191377
DOI:10.1007/s10115-026-02815-0