.png){kind=icon}
.png){kind=icon}
Return
Temporal fuzzy utility-based data analysis on data streams
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.127836.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Temporal fuzzy utility-based data analysis techniques, with their focus on temporal factors in data and the readability of results, provide interesting patterns useful for making rational decisions or predictions in various types of intelligence systems. However, this can be a challenging task considering the massive real-world data streams that are continuously accumulated. This paper suggests an efficient approach to the discovery of temporal fuzzy utility-based patterns for stream data analysis. It discovers interesting patterns with high utility by incorporating linguistic terms through the fuzzy set theory and taking the temporal factors of patterns into account. To optimize performance in data stream environments, the proposed approach introduces an efficient list structure for searching patterns and reduces the search space through novel pruning strategies. Furthermore, the designed technique conducts one scan of the inserted data stream, which is necessary in such scenarios. Thorough experiments on synthetic and real datasets demonstrate that the proposed method excels over the latest algorithms for scalability, runtime, and memory. Compared to the state-of-the-art, the proposed method is up to an order of magnitude faster on real datasets and scales over 1.5 times more efficiently with increasing data size. Further experiments on accuracy and significance show the reliability of the discovered patterns and the effectiveness of the introduced pruning strategies.
Keywords:
Stream data analysis
Temporal fuzzy
Pattern analysis
High utility pattern
Data streams
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W
Organization
No organization information available