Efficient Pattern Mining from Temporal Data Through Correlation based Periodicity Detection

Download Now
Provided by: International Journal of Modern Engineering Research (IJMER)
Topic: Data Management
Format: PDF
Periodicity mining captures the evolution of a data value over time. The existing periodicity mining process is text-based which can be applied only to text data. The paper proposed deals with the periodic patterns in multimedia data which includes text as well as audio and images. Multimedia data such as digital images and audio can be treated as temporal values, since a timestamp is implicitly attached to every instant of the signal. A cross correlation based approach is proposed for periodic mining of multimedia data which has its main application in pattern recognition.
Download Now

Find By Topic