traintraintrain: Gaussian process regression

In weichunliao/baeirGPR: Traing a Gaussian Process Regression and Conduct Predictions.

Description

Training a gaussian process regression model

Usage

traintraintrain(train_x, train_y, pred_method = "sr",
  kname = "gaussiandotrel", ktheta = NULL, kbetainv = NULL, ncpu = -1,
  srsize = NULL, clus_size = NULL)

Arguments

train_x	Matrix; the features of training data set.
train_y	Matrix; y.
pred_method	String; Set the model training approach. cg_direct_lm: conjugate gradient decent with no preconditioning usebigK: use built-in matrix inversion local_gpr: use Local GPR method sr: subset regressors
kname	String; the name of kernel; default value is 'gaussiandotrel'.
ktheta	Numeric vector; store kernel parameter; should be provided when tune_param is FALSE.
kbetainv	Numeric; store kernel parameter betainv; shuld be provided when tune_param is FALSE.
ncpu	Integer; the number of thread to be used; set to -1 to use all threads; default value is -1.
srsize	Positive integer; the size of subtraining dataset; should be provided when pred_method = 'sr'
clus_size	Positive integer; parameter for local_gpr: set the cluster size; should be provided when pred_method = 'local_gpr'.
tsize	Positive integer; parameter used for model tuning (tune_param = TRUE), only tsize of training points will be used for model tuning

Value

return a list having four objects: alpha kparam train_x pred_method

weichunliao/baeirGPR documentation built on May 18, 2019, 9:15 p.m.