Are data structures that use very little space, can be any data structures that are substantively smaller than their base data sets. The design and analysis of effective synopsis data structures offer many algorithmic challenges.
Published in Chapter:
Clustering Algorithms for Data Streams
Christos Makris (University of Patras, Greece) and Nikos Tsirakis (University of Patras, Greece)
Copyright: © 2009
|Pages: 6
DOI: 10.4018/978-1-60566-026-4.ch092
Abstract
The World Wide Web has rapidly become the dominant Internet tool which has overwhelmed us with a combination of rich hypertext information, multimedia data and various resources of dynamic information. This evolution in conjunction with the immense amount of available information imposes the need of new computational methods and techniques in order to provide, in a systematical way, useful information among billions of Web pages. In other words, this situation poses great challenges for providing knowledge from Web-based information. The area of data mining has arisen over the last decade to address this type of issues. There are many methods, techniques and algorithms that accomplish different tasks in this area. All these efforts examine the data and try to find a model that fits to their characteristics in order to examine them. Data can be either typical information from files, databases and so forth, or with the form of a stream. Streams constitute a data model where information is an undifferentiated, byte-by-byte flow that passes over the time. The area of algorithms for processing data streams and associated applications has become an emerging area of interest, especially when all this is done over the Web. Generally, there are many data mining functions (Tan, Steinbach, & Kumar, 2006) that can be applied in data streams. Among them one can discriminate clustering, which belongs to the descriptive data mining models. Clustering is a useful and ubiquitous tool in data analysis.