Some refinements of rough -means clustering
โ Scribed by Georg Peters
- Publisher
- Elsevier Science
- Year
- 2006
- Tongue
- English
- Weight
- 281 KB
- Volume
- 39
- Category
- Article
- ISSN
- 0031-3203
No coin nor oath required. For personal study only.
โฆ Synopsis
Lingras et al. proposed a rough cluster algorithm and successfully applied it to web mining. In this paper we analyze their algorithm with respect to its objective function, numerical stability, the stability of the clusters and others. Based on this analysis a refined rough cluster algorithm is presented. The refined algorithm is applied to synthetic, forest and microarray gene expression data.
๐ SIMILAR VOLUMES
R.D. Reiss's (1980) proposal of quasi-quantiles, i.e. certain kernel estimators of population quantiles Q(u), is extended using a mixing function h(u) and an asymmetric bandwidth function defined by endpoints ~,(u) and 6(u). The second order terms in the expansions of the expected value and variance
This paper contains a link between Probability Proportional to Size (PPS) sampling and interpolations of the classical Jensen inequality. We show that these interpolating inequalities are, in fact, special cases of the conditional Jensen inequality when applied over an appropriate probability space.