Alternate data clustering for fast pattern matching in stream time series data

Abstract

Stream time series retrieval has been a major area of study due to its vast application in various fields like weather forecasting, multimedia data retrieval and huge data analysis. Presently, there is a demand for stream data processing, high speed searching and quick response. In this paper, we use a alternate data cluster or segment mean method for stream time series data, where the data is pruned with a computational cost of O(log w). This approach can be used for both static and dynamic stream data processing. The results obtained are the better than the existing algorithms. © 2012 ICST Institute for Computer Science, Social Informatics and Telecommunications Engineering.