The percolation method for an efficient grouping of data

Trémolières, Raymond

doi:10.1016/0031-3203(79)90035-9

Cited by 19 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Given 1 a multidimensional cloud of points 2 a definition of the density related to each point in that space 3 the value of the scale parameter of this density function (its 'influence radius'), then the resulting density landscape is unique and perfectly well-defined. Tremolières (1979) proposed a general 'percolation' algorithm which is independent of the definition of density or the type of data involved. It aims to rigorously define cluster nuclei, ambivalent frontier-points and atypical points, and works by progressively lowering the density level starting from the densest point and spreading around the nuclei which successively appear.…”

Section: Density Methodsmentioning

confidence: 99%

Paving the way for next generation data-stream clustering: towards a unique and statistically valid cluster structure at any time step

Cuxac

Lelu

Cadot

2011

IJDMMM

View full text Add to dashboard Cite

Abstract:In the domain of data-stream clustering, e.g., dynamic text mining as our application domain, our goal is two-fold and a long term one: 1 at each data input, the resulting cluster structure has to be unique, independent of the order the input vectors are presented 2 this structure has to be meaningful for an expert, e.g., not composed of a huge 'catch-all' cluster in a cloud of tiny specific ones, as is often the case with large sparse data tables.The first preliminary condition is satisfied by our Germen density-mode seeking algorithm, but the relevance of the clusters vis-à-vis expert judgment relies on the definition of a data density, relying itself on the type of graph chosen for embedding the similarities between text inputs. Having already demonstrated the dynamic behaviour of Germen algorithm, we focus here on appending a Monte-Carlo method for extracting statistically valid inter-text links, which looks promising applied both to an excerpt of the Pascal bibliographic database, and to the Reuters-RCV1 news test collection. Though not being a central issue here, the time complexity of our algorithms is eventually discussed. 342P. Cuxac et al.

show abstract

Section: Density Methodsmentioning

confidence: 99%

Paving the way for next generation data-stream clustering: towards a unique and statistically valid cluster structure at any time step

Cuxac

Lelu

Cadot

2011

IJDMMM

View full text Add to dashboard Cite

show abstract

“…Pour essayer de mettre en évidence des liens éventuels entre les marchés boursiers ou, si l'on préfère, pour les regrouper en fonction de leurs comportements tels qu'ils ressortent des variations des cours boursiers mensuels, deux méthodes d'analyse statistique ont été successivement utilisées: l'analyse factorielle en composantes principales [5] et la méthode de Percolation de R. Tremolières [8].…”

Section: Ouunclassified

“…Autant dire que l'analyse factorielle n'est pas spécifiquement une méthode de groupage, et qu'il paraît plus judicieux, toutes les fois que l'analyse inductive rend nécessaire la constitution de groupes homogènes, de recourir à une méthode d'analyse des données statistiques dont l'objectif est le groupage. C'est cette préoccupation qui nous a déjà conduit [1] à utiliser la méthode de « Percolation » de R. Tremolières [8].…”

Section: A) Les Dangers De Déformation Introduits Par Vanalyse Factorunclassified

Comportements comparés des marchés boursiers (1974-1979)

Bensoussan¹

2009

View full text Add to dashboard Cite

This article begins by pointing out, with regard to the study of stock market behaviour, the dangers of distortion inherent in factor analysis in principal components when applied as a method of grouping. As a follow-on to a preceeding contribution dealing with the period 1959-70 this article develops and clarifies the methodological aspects (dangers of factor analysis, utility of percolation method) within the context of the period 1974-79. The advantage of constituting groups from behaviourally homogenous markets, calculated according to the monthly variations in stock exchange rates of the 13 most important markets, is then analysed. The statistical analysis of links between national stock markets requires prudence as regards the use of notions in which the world stock markets are considered as being a whole, the concept of "economic blocks", indeed the same prudence must be exercised when considering the independence or interdependence of markets and the corresponding generalisations, although the long-term tendency would seem to have been modified since 1974 in a direction favourable to the empirical validation of the concept.

show abstract

The Problem of Cluster Analysis

Owsiński

2019

Data Analysis in Bi-Partial Perspective: Clustering and Beyond

View full text Add to dashboard Cite

The percolation method for an efficient grouping of data

Cited by 19 publications

References 7 publications

Paving the way for next generation data-stream clustering: towards a unique and statistically valid cluster structure at any time step

Paving the way for next generation data-stream clustering: towards a unique and statistically valid cluster structure at any time step

Comportements comparés des marchés boursiers (1974-1979)

The Problem of Cluster Analysis

Contact Info

Product

Resources

About