Mining Continuously Changing Data Streams
A data stream is a massive unbounded sequence of data elements continuously generated at a rapid rate. Due to this reason, traditional data mining approach is replaced by systems that are able to mine continuous, high-volume, open-ended data streams as they arrive. Many papers have introduced new algorithms using clustering and classification methods to improve the mining of data streams. We have described the system used to build decision trees using constant memory and constant time per example. The cost effective techniques using output granularity algorithm is discussed. This algorithm is based on load shedding and rate based query optimization. The approach used in this algorithm is data rate adaptation. Data rate adaptation maximizes the output accuracy given available resources like memory, CPU utilization, and battery consumption etc. The data is generated rapidly and is continuously changing; therefore there are many issues that need to be taken care of by mining algorithms. In this paper we will discuss three such most widely used algorithms.
LU Yi-hong WANG Zi-ren HUANG Yan
College of Information Engineering, Zhejiang University of Technologeg Hangzhou, 310032, China University of North Texas, USA
国际会议
北京
英文
2007-08-05（万方平台首次上网日期，不代表论文的发表时间）