.png){kind=icon}
返回
3P-ECLAT: mining partial periodic patterns in columnar temporal databases
.png){kind=icon}
PRE
AI
DOI：10.1007/s10489-023-05172-5.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Partial periodic pattern (3P) mining is a vital data mining technique that aims to discover all interesting patterns that have exhibited partial periodic behavior in temporal databases. Previous studies have primarily focused on identifying 3Ps only in row temporal databases. One can not ignore the existence of 3Ps in columnar temporal databases as many real-world applications, such as Facebook and Adobe, employ them to store their big data. This paper proposes an efficient single database scan algorithm, Partial Periodic Pattern-Equivalence Class Transformation (3P-ECLAT), to identify all 3Ps in a columnar temporal database. The proposed algorithm compresses the given database into a novel list-based data structure and mines it recursively to find all 3Ps. The 3P-ECLAT leverages the downward closure property and depth-first search technique to reduce the search space and the computational cost. Extensive experiments have been conducted on synthetic and real-world databases to demonstrate the efficiency of the 3P-ECLAT algorithm. The memory and runtime results show that 3P-ECLAT outperforms its competitor considerably. Furthermore, 3P-ECLAT is highly scalable and is superior to the previous approach in handling large databases. Finally, to demonstrate the practical utility of our algorithm, we provide two real-world case studies, one on analyzing traffic congestion during disasters and another on identifying the highly polluted areas in Japan.
Keyword:
Data mining
Pattern mining
Partial periodic patterns
Columnar temporal databases
期刊
IF:
3.5
论文数:
7.6K
被引数:
1.7W
机构
引用论文
.png){kind=icon}
Spatial Variation in Sound Frequency Components Across an Urban Area Derived from Mobile Surveys
OAAI
OAAI分享收藏分享收藏
查看更多内容