Internal function for doing simulation using functional lars.

Share:

Description

This is a function built for doing data generation and variable selection using functional lars with different settings and data with different correlation structures.

Usage

1
2
3
4
5
flars_TrainTest(seed=1,nsamples=120,nTrain=80,var_type=c('f','m'),
                VarThreshold0=0.1,SignThreshold0=0.8,cor_type=1:5,
                lasso=TRUE, check = 1,uncorr=T,nVar=8,Discrete_Norm_ID=1:12,
                NoRaw_max=12,raw_max=9,hyper=NULL,RealX=NULL,RealY=NULL,
                dataL=NULL,nCor=0,control=list())

Arguments

seed

Set the seed for random numbers.

nsamples

Sample size of the data to generate.

nTrain

Sample size of the training data.

var_type

Two choices of the variable types. See details for more information.

cor_type

Correlation structures. See details for more information.

VarThreshold0

Threshold for removing variables based on variation explained. See flars for more details.

SignThreshold0

Same as VarThreshold0

lasso

Use lasso modification or not. In other words, can variables selected in the former iterations be removed in the later iterations.

check

Type of lasso check. 1 means variance check, 2 means sign check. check=1 is much better than the other one.

uncorr

If the variables are uncorrelated or not. See details for more information.

nVar

Number of variables to generate.

Discrete_Norm_ID

Which discrete method and which norm to use. 1 to 12.

NoRaw_max

Number of variables to select when not using RDP discretising method.

raw_max

Number of variables to select when using RDP discretising method.

hyper

Hyper parameters used in the Gaussian process. GP is used for building the covariance structure of the functional variables.

RealX

Real data input X.

RealY

Real data input Y.

dataL

Real input data list rather than generate in the function. It should has the same structure as that generated.

nCor

Number of cores to use.

control

List of control items. See fccaGen for more details.

Value

A list of results using different normalization methods and different representation methods for the functional coefficients.