Search PubMed⌕ Search

PubMed · 11272743

Validating cluster assignments.

Abstract

In the absence of a statistically based assessment of cluster assignments of observations in terms of their similarity along selected variables the validity of a set of clusters could be developed if clusters withstand both empirical scrutiny and "make sense" to expert informants.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

D J Ketchen, G T Hult. 2000. Validating cluster assignments.. https://doi.org/10.2466/pr0.2000.87.3f.1057

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

Clustering individuals using INMTD: a novel versatile multi-view embedding framework integrating omics and imaging data.

MOTIVATION: Combining omics and images can lead to a more comprehensive clustering of individuals than classic single-view approaches. Among the various approaches for multi-view clustering, nonnegative matrix tri-factorization (NMTF) and nonnegative Tucker decomposition (NTD) are advantageous in learning low-rank embeddings with promising interpretability. Besides, there is a need to handle unwanted drivers of clusterings (i.e. confounders). RESULTS: In this work, we introduce a novel multi-view clustering method based on NMTF and NTD, named INMTD, which integrates omics and 3D imaging data to derive unconfounded subgroups of individuals. According to the adjusted Rand index, INMTD outperformed other clustering methods on a synthetic dataset with known clusters. In the application to real-life facial-genomic data, INMTD generated biologically relevant embeddings for individuals, genetics, and facial morphology. By removing confounded embedding vectors, we derived an unconfounded clustering with better internal and external quality; the genetic and facial annotations of each derived subgroup highlighted distinctive characteristics. In conclusion, INMTD can effectively integrate omics data and 3D images for unconfounded clustering with biologically meaningful interpretation. AVAILABILITY AND IMPLEMENTATION: INMTD is freely available at https://github.com/ZuqiLi/INMTD.

Cluster Analysis↗

Fuzzy species among recombinogenic bacteria.

BACKGROUND: It is a matter of ongoing debate whether a universal species concept is possible for bacteria. Indeed, it is not clear whether closely related isolates of bacteria typically form discrete genotypic clusters that can be assigned as species. The most challenging test of whether species can be clearly delineated is provided by analysis of large populations of closely-related, highly recombinogenic, bacteria that colonise the same body site. We have used concatenated sequences of seven house-keeping loci from 770 strains of 11 named Neisseria species, and phylogenetic trees, to investigate whether genotypic clusters can be resolved among these recombinogenic bacteria and, if so, the extent to which they correspond to named species. RESULTS: Alleles at individual loci were widely distributed among the named species but this distorting effect of recombination was largely buffered by using concatenated sequences, which resolved clusters corresponding to the three species most numerous in the sample, N. meningitidis, N. lactamica and N. gonorrhoeae. A few isolates arose from the branch that separated N. meningitidis from N. lactamica leading us to describe these species as 'fuzzy'. CONCLUSION: A multilocus approach using large samples of closely related isolates delineates species even in the highly recombinogenic human Neisseria where individual loci are inadequate for the task. This approach should be applied by taxonomists to large samples of other groups of closely-related bacteria, and especially to those where species delineation has historically been difficult, to determine whether genotypic clusters can be delineated, and to guide the definition of species.

Cluster Analysis↗

Finite-size effect in the Eguíluz and Zimmermann model of herd formation and information transmission.

The Eguíluz and Zimmermann model of information transmission and herd formation in a financial market is studied analytically. Starting from a formal description on the rate of change of the system from one partition of agents in the system to another, a mean-field theory is systematically developed. The validity of the mean-field theory is carefully studied against fluctuations. When the number of agents N is sufficiently large and the probability of making a transaction a<<1/N ln N, finite-size effect is found to be significant. In this case, the system has a large probability of becoming a single cluster containing all the agents. For small clusters of agents, the cluster size distribution still obeys a power law but with a much reduced magnitude. The exponent is found to be modified to the value of -3 by the fluctuation effects from the value of -5/2 in the mean-field theory.

Cluster Analysis↗