This function takes a definition of weight transformation limits and corresponding minimum and maximum numbers of end-members to model all end-member scenarios in accordance with these parameters. Based on the output the user can decide on robust end-members.
model.EM(X, q, l, classunits, plot = TRUE, col.q = TRUE, bw, ...)Numeric matrix, input data set with m samples (rows)
and n variables (columns).
Numeric matrix, definitions of minimum and maximum number
of end-members (cf. get.q()), required.
Numeric vector, weight transformation limit values,
corresponding to the matrix q, required.
Numeric vector, optional class units
(e.g. micrometers or phi-units) of the same length as columns of X.
Logical scalar, option to plot the results (cf.
details for explanations), default is TRUE.
Logical scalar, option to colour end-member loadings by
the number of end-members which were used to create the model realisation,
default is TRUE.
Numeric scalar, optional manual setting of the kde
bandwidth. By default, bw is calculated as 1 percent of the number of
grain-size classes.
Further arguments passed to the function.
List object with all modelled end-members, each described by
input parameters, mode position, quality measures and value distributions.
The plot output is an overlay of several data. The coloured lines in the
background are end-member loadings (number noted in the plot title),
resulting from all possible model scenarios. If col.q == TRUE they
are coloured according to the number of end-members with which the model
was generated. This colour scheme allows to depict end-members that emerge
for model realisations with specific number of end-members. The thick
black line is a kernel density estimate curve, generated from the mode
positions of all end-members. The kernel bandwidth is set to 1 percent of
the number of grain-size classes of the input data set, which gave useful
results for most of our test data sets. The cumulaitve dot-line-plot is a
further visualisation of end-member mode positions. The function is a
modified wrapper function for the function test.robustness().
Dietze E, Hartmann K, Diekmann B, IJmker J, Lehmkuhl F, Opitz S, Stauch G, Wuennemann B, Borchers A. 2012. An end-member algorithm for deciphering modern detrital processes from lake sediments of Lake Donggi Cona, NE Tibetan Plateau, China. Sedimentary Geology 243-244: 169-180.
# NOT RUN {
## load example data set
data(example_X)
## define input parameters
l <- c(0, 0.05, 0.10)
q <- cbind(c(2, 2, 3), c(5, 6, 4))
## infer l-vector
em_pot <- model.EM(X = X, q = q, l = l)
# }
Run the code above in your browser using DataLab