.png){kind=icon}
.png){kind=icon}
Return
Regularity-driven pattern extraction and analysis approach by the pre-pruning technique without pattern loss
.png){kind=icon}
PRE
AI
DOI：10.1016/j.future.2024.107670.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Pattern analysis is responsible for a significant role in data extraction as we enter the era of big data, providing valuable information. Regular patterns, which are temporally consistent patterns in transactional data, offer significant and intelligent insights in various areas. Temporal regularity in a regular pattern allows analyzing and recognizing noteworthy knowledge that appear recurrently from sensor data of the Internet of Things (IoT), such as medical sensors. Regular pattern analysis has been studied to find regular patterns using the temporal regularity of pattern occurrences in huge amounts of data. Recently, the development of tree-based regular pattern analysis algorithms has progressed. However, this tree-based data structure takes long time and require huge memory. For this reason, we suggest a novel list-based data structure and present an algorithm using the proposed structure to discover regular patterns on transactional data efficiently. Our algorithm extracts the exact regular pattern results while spending less time and memory than the existing methods because the proposed structure stores information from the data in minimal structure expression. In performance evaluation, we compare our method to existing methods. These performance tests demonstrate our algorithm outperforms other methods in runtime, memory efficiency, and scalability. Moreover, the accuracy test shows that our method discovers regular patterns accurately. The significance test demonstrates the advantage of regularity, which the proposed algorithm considers. Furthermore, the implications of the proposed algorithm are discussed with concrete applications. In addition, we discuss methods for handle huge amounts of data in the real world.
Keywords:
Regularity driven pattern analysis
Pre-prunning technique
Transactional data
Journal
F
IF:
6.1
Papers:
6.8K
Citations:
2.3W