RT Journal Article
T1 Cluster identification using projections
A1 Peña, Daniel
A1 Prieto, Francisco J.
AB This artiele describes a procedure to identify elusters in multivariate data using information obtained from the univariate projectionsof the sample data onto certain directions. The directions are chosen as those that minimize and maximize the kurtosis coefficlent ofthe projected data. It is shown that, under certain conditions, these directions provide the largest separatlOn for the dlfferent clusters.The projected univariate data are used to group the observations according to the values of the gaps or spacings between consecutive-orderedobservations. These groupings are then combined over all projection directions. The behavior of the method is tested on severalexamples, and compared to k-means, MCLUST, and the procedure proposed by Jones and Sibson in 1987. The proposed algonthm isiterative, affine equivariant, flexible, robust to outliers, fast to implement, and seems to work well in practice
PB American Statistical Association
SN 0162-1459
YR 2001
FD 2001-12
LK https://hdl.handle.net/10016/15524
UL https://hdl.handle.net/10016/15524
LA eng
NO This research was supported by Spanish grant BEC2000-0167
DS e-Archivo
RD 1 may. 2024