Abstract:
Discovering interesting patterns or substructures in data streams
is an important challenge in data mining. Clustering algorithm are very
often applied to identify substructures, although they are designed to
partition a data set. Another problem of clustering algorithms is that most
of them are not designed for data streams. They assume that the data set to
be analysed is already complete and will not be extended by new data. This
paper discusses an extension of an algorithm that uses ideas from cluster
analysis, but was designed to identify single clusters in large data sets
without the necessity to partition the whole data set into clusters. The new
extended version of this algorithm can applied to stream data and is able to
identify new clusters in an incoming data stream. As a case study weather
data are used