fitSpectra-class: Create a list with covariance matrices of the spectra and...

In asmitapoddar/BayesSentinel: Bayes Classification of Spetrometric Data

Description

Return the covariance matrices

Usage

fitSpectra(...)

Arguments

...	any paramaters to be input into the function

Slots

m

spectroscopic data

modelname

name of model to be used for calculating the covariance matrix. Available models are "full", "parsimonious". Default is "full".

spectra

type of spectra. Available models are "diag", "unknown" and "kernel". Default is "diag".

time

type of time. Available models are "diag", "unknown" and "kernel". Default is "diag".

kerneltypeSpectra

kernel to be used for covariance matrix of spectra Available kernels are "epanechnikov", "gaussian", "exponential", "uniform", "quadratic", "circular", "triangular", "rational quadratic", "inverse multiquadratic". Default is "exponential".

kerneltypeTime

kernel to be used for covariance matrix of time Available kernels are "epanechnikov", "gaussian", "exponential", "uniform", "quadratic", "circular", "triangular", "rational quadratic", "inverse multiquadratic". Default is "exponential".

h

used for kernel calculation

s

correction limit paramater for flip flop algorithm

lambdaS

regularisation for spectra for flip flop algorithm

lambdaT

regularisation for spectra for flip flop algorithm

validation

to optimize lambda in case of th model is : M = parsimonious, S=unknown, T=unknow

listLambdaS

list of lambdaS used in prediction in case validation is TRUE

listLambdaT

list of lambdaT used in prediction in case validation is TRUE

model

use in prediction in case of validation is TRUE

covMat

returning the covariance matrx

Author(s)

Asmita Poddar & Florent Latimier

asmitapoddar/BayesSentinel documentation built on May 10, 2019, 1:18 a.m.