Exploring Multivariate Data with the Forward Search

Atkinson, Anthony C.; Riani, Marco; Cerioli, Andrea

doi:10.1007/978-0-387-21840-3

Cited by 168 publications

(152 citation statements)

References 0 publications

Supporting

Mentioning

149

Contrasting

Unclassified

Order By: Relevance

“…It is the purpose of the present paper to extend and apply robust cluster analysis using the forward search as intro duced in Chapter. 7 of Atkinson et al (2004). This graphics-rich robust approach to clustering uses the data to identify the number of clusters, to confirm cluster mem bership and to detect outlying observations that do not belong to any cluster.…”

Section: Introductionmentioning

confidence: 99%

Robust Clustering for Performance Evaluation

Atkinson

Riani²,

Cerioli³

2009

Data Analysis and Classification

Self Cite

View full text Add to dashboard Cite

The evaluation of the effectiveness of organisations can be aided by the use of cluster analysis, suggesting and clarifying differences in structure between successful and failing organisations. Unfortunately, traditional methods of cluster analysis are highly sensitive to the presence of atypical observations and departures from normality. We describe a form of robust clustering using the forward search that allows the data to determine the number of clusters and so allows for outliers. An example is given of the successful clustering of customers of a bank into groups that are decidedly non-normal.

show abstract

Section: Introductionmentioning

confidence: 99%

Robust Clustering for Performance Evaluation

Atkinson

Riani²,

Cerioli³

2009

Data Analysis and Classification

Self Cite

View full text Add to dashboard Cite

show abstract

“…Again, looking at the data from different perspectives and inspecting the composition of S(m) just before the major peaks in the forward plots of d min (m) lead in the end to the identification of the whole group structure. Procedures similar to those described in [ARC04] and plots like those in Figure 4 could be used to confirm this structure and to explore borderline units.…”

Section: Simulation Resultsmentioning

confidence: 99%

“…These units might also be difficult to detect using multivariate techniques with a high breakdown point, as is shown in [ARC04] in the case of bivariate continuous populations. Alternative clustering methods that try to reduce the effect of outliers are described by [CAGM97], [FR02] and [H03], but most of them are not easily extended to deal with nominal categorical variables.…”

Section: Outlier Detectionmentioning

confidence: 99%

“…The algorithm suggested in this paper shares the main properties of the forward search approach described in [ARC04], including computational simplicity and effectiveness of its graphical displays. It can be implemented by suitable modification of the software Rfwdmv, an R package specifically devised to perform the forward search for the analysis of continuous multivariate observations.…”

Section: Computational Issuesmentioning

confidence: 99%

“…The forward search is a powerful general method for detecting unidentified subsets and multiple masked outliers and for determining their effect on models fitted to the data. The search for multivariate data, including cluster analysis with quantitative variables, is given book length treatment by Atkinson, Riani and Cerioli [ARC04]. It is our purpose to extend their robust clustering technique to cope with non-numeric attributes.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Robust classification with categorical variables

Cerioli

Riani

Atkinson

Compstat 2006 - Proceedings in Computational Statistics

Self Cite

View full text Add to dashboard Cite

Summary.The forward search provides a powerful and computationally simple approach for the robust analysis of multivariate data. In this paper we suggest a new forward search algorithm for clustering multivariate categorical observations. Classification based on categorical information poses a number of challenging issues that are addressed by our algorithm. These include selection of the number of groups, identification of outliers and stability of the suggested solution. The performance of the algorithm is shown with both simulated and real examples.

show abstract

Outliers

Aelst¹

2012

Encyclopedia of Environmetrics

View full text Add to dashboard Cite

Datasets frequently contain outliers which are observations that strongly deviate from the pattern in the majority of the data. Detecting outliers and measuring their influence on an estimation procedure is not trivial. In one or two dimensions graphical inspection of the data may be helpful, but in higher dimensions more reliable methods are needed. We discuss robust estimation to obtain estimates that can resist the influence of outliers and related outlier diagnostics to identify the outliers in a dataset.

show abstract

Exploring Multivariate Data with the Forward Search

Cited by 168 publications

References 0 publications

Robust Clustering for Performance Evaluation

Robust Clustering for Performance Evaluation

Robust classification with categorical variables

Outliers

Contact Info

Product

Resources

About