Download now Free registration required
Clustering algorithms partition a set of objects into clusters based on similarity of objects. A clustering is therefore a set of clusters, where similar objects are placed in the same cluster. Various clustering methods have been studied, for example, partitional, hierarchical, density-based, graph-based, neural network, fuzzy, compression-based, and consensus clusterings. In database applications, existing clustering algorithms are often needed to be adapted for two reasons. First, these algorithms assume that the entire set of data can fit in the memory to achieve efficiency. In practice, however, the size of the data set may be several magnitudes larger than that of the memory. Second, these algorithms tend to ignore I/O cost during data processing.
- Format: PDF
- Size: 271.9 KB