coupures.Mp {SARP.compo}

Finding cut-offs for graph disjonctions

Description

These functions detect the experimental cut-offs to create distinct subgraphs, and propose adapted graphical representation.

Usage

coupures.Mp( Mp )

## S3 method for class 'Coupures'
plot(x, seuil.p = 0.05, en.log = TRUE,
                        xlab = "Seuil de p", ylab = "Nombre de composantes",
                        col.trait = "black", lwd.trait = 1, lty.trait = 1,
                        col.seuil = "red"  , lwd.seuil = 1, lty.seuil = 1,
                        pch.fin = 19, cex.fin = 1, col.fin ="darkgreen",
                        pch.deb = ")", cex.deb = 1, col.deb = "darkgreen",
                        ...)

Arguments

Mp

A square, symmetric matrix containing p-values. Element in row i and line j should contain the p-value for testing the \frac{i}{j} ratio. The diagonal is ignored.

x

The set of critical values, as obtained by coupures.Mp

seuil.p

Selected cut-off for analysis. Can also be a SARPcompo.H0 object, as returned by choisir.seuil, in which case the bounds of the confidence interval are also drawn, with dashed lines by default.

en.log

If TRUE, the p-values axis uses a decimal logarithm scale. It may help visualisation of the small critical p-values.

xlab, ylab

Legends for the axes

col.trait, lwd.trait, lty.trait, pch.fin, cex.fin, col.fin, pch.deb, cex.deb, col.deb

Graphical parameters for drawing the number of components in function of the cut-off. ‘trait’ refers to the function itself, ‘deb’ to the first point of a region of constant components number (that does not belong to it: the function is right-discontinuous) and ‘fin’ to the last point of this region (that belongs to it)

col.seuil, lwd.seuil, lty.seuil

Graphical parameters for drawing the analysis cut-off

...

Additionnal parameters for plot, which is used internally.

Details

By increasing the cut-off from 0 to 1, more and more edges between nodes are removed, and disjoint subgraphs appear. This function detects in a matrix of p-values which are the “critical” ones, that is the one for which the number of components changes.

Because the edge removal is defined by p < cut-off, the cut-off returned for a given number of components is to be understand as the maximal one that gives this number of components.

The plot method allows to visualize the evolution of the number of components with the cut-off, and writes the critical cut-off values.

Value

The coupures.Mp function returns a data.frame with additionnal class ‘Coupures’. It contains three columns: one with the p-value cut-offs, one with the opposite of their decimal logarithm and one with the number of components when using exactly this cut-off. The additionnal class allows to provide a plot method.

Author(s)

Emmanuel Curis ([e-mail])

See Also

creer.Mp to create a matrix of p-values for all possible ratios of a compositional vector.

grf.Mp to convert such a matrix to a graph, once a cut-off is selected.

arbre.Mp to convert such a matrix to a classification tree of the components of the compositional vector.

Examples

   # load the potery data set
   data( poteries )

   # Compute one-way ANOVA p-values for all ratios in this data set   
   Mp <- creer.Mp( poteries, c( 'Al', 'Na', 'Fe', 'Ca', 'Mg' ),
                   f.p = anva1.fpc, v.X = 'Site' )

   # Where would be the cut-offs?
   seuils <- coupures.Mp( Mp )
   seuils

   # Drawing this, in log10 scale
   plot( seuils, en.log = TRUE )