Title :
Discovering Temporal Similarity Pattern Based on Metamorphosis Data
Author_Institution :
Comput. & Commun. Dept., Hunan Inst. of Eng., Xiangtan, China
Abstract :
This paper discusses a metamorphosis method for temporal data mining. We propose definition of temporal type and time granularity to divide into segments for time, then build an event temporal space which may describe something change of data over time. We give the conception of metamorphosis data in event temporal types space so that it is easier to discover valuable knowledge, but it is not clear to do it if original data are not metamorphosis. By metamorphosis data, we study a problem of knowledge discovery of temporal similarity pattern based on the temporal type is given. A method based on self-organizing map (SOM) to find similarity pattern is proposed.
Keywords :
data mining; self-organising feature maps; event temporal types space; knowledge discovery; metamorphosis data; self-organizing map; temporal data mining; temporal similarity pattern discovery; Association rules; Cancer; Computational intelligence; Computer security; Data engineering; Data mining; Data security; Medical diagnostic imaging; Monitoring; Time series analysis;
Conference_Titel :
Computational Intelligence and Security, 2009. CIS '09. International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-5411-2
DOI :
10.1109/CIS.2009.71
