phytoclass (version 2.3.1)

NNLS_MF_Final: Perform matrix factorisation for phytoplankton pigments and pigments ratios

Description

Performs the non-negative matrix factorisation for given phytoplankton pigments and pigment ratios, to attain an estimate of phytoplankton class abundances.

Usage

NNLS_MF_Final(Fn, S, S_Chl, S_weights, S_dvChl = NULL)

Value

A list containing the following elements:

F matrix

The normalized F matrix of pigment ratios

RMSE

Root mean square error of the fit

condition number

Condition number of Fn %*% t(S)

Class abundances

Data frame of phytoplankton class abundances

Figure

Plot of the results

MAE

Mean absolute error for each pigment

Error

Residual error matrix

Arguments

Fn

F matrix with pigment ratios for each phytoplankton class

S

Sample data matrix of pigment measurements

S_Chl

Vector of chlorophyll a concentrations for each sample

S_weights

Vector of weights for each pigment

S_dvChl

Optional vector of divinyl chlorophyll concentrations for Prochlorococcus

Details

Unlike NNLS_ML(), it also removes any weighting and normalisation, and also multiplies relative abundances by chlorophyll values to determine the biomass of phytoplankton groups.

Examples

 Fmat <- as.matrix(phytoclass::Fm)
 S <- as.matrix(phytoclass::Sm)
 S_weights <- as.numeric(phytoclass:::Bounded_weights(S))
 S_Chl <- S[, ncol(S)]
 # Run NNLS_MF_Final
 result <- phytoclass:::NNLS_MF_Final(Fmat, S, S_Chl, S_weights)