finemeshed3d: finemeshed3d

In SpiceFP: Sparse Method to Identify Joint Effects of Functional Predictors

finemeshed3d

Description

Function that helps to transform a vector into a 3 dimensional array (with a fine mesh). In the implementation of the spiceFP approach, it allows to transform matrices of coefficients having different dimensions into matrices of the same dimension in order to perform arithmetic operations. In practice, the 3d array to be transformed is associated with a contingency table, which implies numerical variables for which classes have been created.

Usage

finemeshed3d(
  x,
  n.breaks1 = 10,
  n.breaks2 = 1000,
  n.breaks3 = 500,
  round.breaks1 = 9,
  round.breaks2 = 9,
  round.breaks3 = 9
)

Arguments

x	vector or one column matrix to scale. This vector comes from the vectorization of the 3d array to be transformed. x is named using the concatenation of the names of the dimension of the array to be transformed, as shown in the example below.
n.breaks1	integer. Number of breaks needed for the first variable The variable for which classes are in first position when constructing x's names is the first variable.
n.breaks2	integer. Number of breaks needed for the second variable. The variable for which classes are in second position when constructing x's names is the second variable.
n.breaks3	integer. Number of breaks needed for the third variable. The variable for which classes are in third position when constructing x's names is the third variable.
round.breaks1	integer. Number of decimals for breaks of the first variable.
round.breaks2	integer. Number of decimals for breaks of the second variable.
round.breaks3	integer. Number of decimals for breaks of the third variable.

Details

This function is designed to return a 3d fine meshed array and breaks associated. In order to obtain a fine mesh, a high number of breaks must be fixed.

Value

Returns:

finemeshed.array

Array of dimension n.breaks1 x n.breaks2 x n.breaks3. The dimension names of finemeshed.array are the breaks created from each variable and the associated n.breaks. Each value of finemeshed.array is equal to the value of x indexed by the classes containing the row and column names of finemeshed.array

finemeshed.values1

First variable breaks

finemeshed.values2

Second variable breaks

finemeshed.values3

Third variable breaks

Examples

set.seed(4)
count_table<-hist_3d(x = rnorm(1000),
                     y = rnorm( 1000,5,0.1),
                     z = rnorm( 1000,2,1),
                     breaks_x = seq(-4, 4, by =1),
                     breaks_y = seq(2, 8, by =1),
                     breaks_z = seq(-3, 6, by =1))$Hist.Values

df.x<-as.data.frame.table(count_table)
x<-df.x$Freq
names(x)<-paste0(df.x$Var1,"_",df.x$Var2,"_",df.x$Var3)

res.fm3d<- finemeshed3d(x,10,50,100)
dim(res.fm3d$finemeshed.array)

SpiceFP documentation built on June 7, 2023, 5:55 p.m.