We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Linkage Pattern Mining Method for Multiple Sequential Data with Noise.
- Authors
Saerom Lee; Takahiro Miura; Yusuke Okubo; Yoshifumi Okada
- Abstract
Linkage pattern mining is a data mining technique that finds frequent patterns that appear repeatedly across multiple sequential data. This technique does not assume similarity or correlation between the frequent patterns in a linkage pattern; thus, it is expected to be a promising approach for discovering causal association among events in multiple sensor data, such as physiological signals in different regions and crustal movements at different points. However, existing methods have focused only on detecting linkage patterns without noise/fluctuations in sequential data. This study's objective is to develop a new noise-robust linkage pattern mining method. The proposed method excludes pseudo patterns derived from noise using closed itemset mining from interval graphs regarding frequent patterns such that only noiseless and maximal linkage patterns are extracted. The proposed method is applied to artificial sequential datasets with embedded linkage patterns. Experimental results show that this method can adequately detect embedded linkage patterns without noise and previously undetectable embedded linkage patterns with noise.
- Subjects
SEQUENTIAL pattern mining; NOISE; DATA analysis; LIAISON theory (Mathematics); GRAPH theory
- Publication
IAENG International Journal of Computer Science, 2015, Vol 42, Issue 4, p361
- ISSN
1819-656X
- Publication type
Article