A Method for Comparing Two Hierarchical Clusterings

Fowlkes, Edward B.; Mallows, C. L.

doi:10.1080/01621459.1983.10478008

Cited by 1,064 publications

(300 citation statements)

References 5 publications

Supporting

Mentioning

292

Contrasting

Unclassified

Order By: Relevance

“…The FM index [27] was found to be superior to the other indices when reference datasets are generated under the uniformity hypothesis (data not shown).…”

Section: Resultsmentioning

confidence: 99%

A prediction-based resampling method for estimating the number of clusters in a dataset

2002

View full text Add to dashboard Cite

show abstract

“…The FM index [27] was found to be superior to the other indices when reference datasets are generated under the uniformity hypothesis (data not shown).…”

Section: Resultsmentioning

confidence: 99%

A prediction-based resampling method for estimating the number of clusters in a dataset

2002

View full text Add to dashboard Cite

show abstract

“…We then rerun the clustering algorithm based on the subsampling a fraction of the samples and group the subsamples into k clusters. We then compute a similarity index of the subsamples, the correlation coefficient between the clusters for the resampled data with those for the original data based on the definition given by Fowlkes and Mallows [18]. We repeat this several times to get a histogram of correlation coefficient values.…”

Section: Systems and Methodsmentioning

confidence: 99%

Cluster stability scores for microarray data in cancer studies

Smolkin

Ghosh

2003

BMC Bioinformatics

View full text Add to dashboard Cite

BackgroundA potential benefit of profiling of tissue samples using microarrays is the generation of molecular fingerprints that will define subtypes of disease. Hierarchical clustering has been the primary analytical tool used to define disease subtypes from microarray experiments in cancer settings. Assessing cluster reliability poses a major complication in analyzing output from clustering procedures. While most work has focused on estimating the number of clusters in a dataset, the question of stability of individual-level clusters has not been addressed.ResultsWe address this problem by developing cluster stability scores using subsampling techniques. These scores exploit the redundancy in biologically discriminatory information on the chip. Our approach is generic and can be used with any clustering method. We propose procedures for calculating cluster stability scores for situations involving both known and unknown numbers of clusters. We also develop cluster-size adjusted stability scores. The method is illustrated by application to data three cancer studies; one involving childhood cancers, the second involving B-cell lymphoma, and the final is from a malignant melanoma study.AvailabilityCode implementing the proposed analytic method can be obtained at the second author's website.

show abstract

“…The number of cluster is fixed by using some prior working hypotheses, in which data are grouped basing on specific investigation criteria. The clustering achieved are finally compared with the ideal grouping using a quantitative index, the Fowlkes-Mallows index (FMI) (Fowlkes and Mallows, 1983). It is defined as:…”

Section: Multivariate Analysis Methodsmentioning

confidence: 99%

“…Clustering is carried out by using the selected set of features and evaluating the level of clusterization. Different clustering solutions have been tested according to specific strategies on the kind of factors to investigate on (high quality brands vs. low quality brands, or geographic origin are two possible examples) and adopting a measure of cluster similarity, the Fowlkes-Mallows index (Fowlkes and Mallows, 1983), to evaluate the goodness of the clustering achieved.…”

Section: Feature Extraction Which Includesmentioning

confidence: 99%

Digital Fingerprinting of Coffee Blending by Sensitive Crystallization

Guccione¹,

Galati²,

Pusceddu³

et al. 2017

JFR

View full text Add to dashboard Cite

The authenticity and quality of productions is an area of priority interest that involves safety of consumers and potential economic damages deriving from frauds on origin, adulteration and labeling of products. Several investigation techniques are currently used to characterize food matrices from physical-chemical-biological point of view using different methods in order to limit possible adulterations. In this work, we have developed an experimental and computational framework to improve the potentialities of sensitive crystallization: an experimental technique known since 1936, but never used for quantitative assessment of food quality. As a test case, it has been applied to investigate the geographical traceability and quality of coffee samples. An extensive statistical analysis associated with a careful choice of advanced image descriptors allows gathering quantitative information about the samples, which can constitute a digital fingerprint of their composition. With this new tool we are able to distinguish with blind tests high-quality coffee brands from low-quality mixtures, different coffee species, green from toasted condition of beans and, to a lesser extent, the macro-geographical provenience. A powder X-ray diffraction analysis reinforces the results obtained by sensitive crystallization for the case where crystalline domains are present in the coffee sample.

show abstract

A Method for Comparing Two Hierarchical Clusterings

Cited by 1,064 publications

References 5 publications

A prediction-based resampling method for estimating the number of clusters in a dataset

A prediction-based resampling method for estimating the number of clusters in a dataset

Cluster stability scores for microarray data in cancer studies

Digital Fingerprinting of Coffee Blending by Sensitive Crystallization

Contact Info

Product

Resources

About