.png){kind=icon}
Return
Approximate erasable pattern discovery and analytics on stream data
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2025.113161.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Erasable pattern mining, which finds patterns that can be removed by having a low gain value, minimizes losses by finding products that have less profit to overcome financial crises occurring in industrial fields such as manufacturing factories. When using data in the real world, there are errors of various sizes in the data, such as measurement errors, communication errors, or noise. The results from these environments are not accurate and the defects in the pattern extracted in this way may cause long-lasting effects and significant losses in the industrial field. Accordingly, we propose a mining strategy that extracts reliable and robust erasable patterns in any environment. Our proposed algorithm applies the concept of approximate factor to consider an error range of data and the number of transactions. It also extracts erasable patterns quickly and efficiently by considering characteristics of data streams generated and accumulated in real time. In this paper, approximate erasable pattern mining is performed from the data stream to extract reliable erasable patterns even in an environment where errors may occur. The performance evaluations are conducted using four real datasets that have various characteristics and two synthetic dataset groups. The results of the evaluations demonstrate our approach is faster the maximum four times over regarding runtime and generally efficient regarding memory usage.
Keywords:
Data mining
Approximation
Erasable pattern
Error tolerance
Stream data
Journal
K
IF:
7.6
Papers:
1.3W
Citations:
4.5W
Organization
No organization information available
Cited Papers
Approximation of Probabilistic Maximal Frequent Itemset Mining Over Uncertain Sensed Data.png)
IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSave
ShareSaveShareSaveA decision-support model through online reviews: Consumer preference analysis and product ranking
PREAI
ShareSave
View more