In gusart/tuckerR_mmgg: Three-Mode Principal Components Analysis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "./tools/"
)

tuckerR.mmgg

This package performs Three-Mode Principal Components using Tuckers Models and plot interactive Biplot.Some experiment design generated three-way or three-mode data, repeated observations of a set of attributes for a set of individuals in different conditions. The information was displayed in a three-dimensional array, and the structure of the data was explored using Three-Mode Principal Component Analysis, the Tucker-2 Model.

Installation

You can install tuckerR.mmgg from github with:

# install.packages("devtools")
devtools::install_github("gusart/tuckerR_mmgg")

Also you can install from CRAN as a packages

install.packages("tuckerR.mmgg"")

Important contribution of this package

The most important contribution of this package are the interactive biplot graphics and the application of the diffit() function to find the best combination of components to retain.

Example

The data are part of collection of local populations corresponding to different races that are conserved in the Active Germplasm Bank of INTA Pergamino Experimental Station, Argentina.A data frame with 10 characteristics of 31 maize populations in two different conditions corresponding to production areas of Buenos Aires. Since the variables are repeated in both places the data frame has a total of 20 variables, 10 for an environment and evaluated them in the other conditions.

Compute the number of components to retain using diffit methods

The diffit method is used to apply when we need to know the axis number to be gathered in the P mode, and Q mode. The third mode, K it is related to the environment numbers. The diffit method consist on fitting each value with the Tuckle algorithm.

library(tuckerR.mmgg)
data(maize_pop,package = "tuckerR.mmgg")
dif_sal <- diffit(maize_pop,amb=2)
print(dif_sal)

In determining the number of components to be retained, the first row must be ignored, since it is the first value in the algorithm. In this case, the best combination is $3-3-2$.

This is a basic example which shows you how to solve a common problem:

data(maize_pop)
output <- tucker2R(maize_pop,amb=2,stand=TRUE,nc1=3,nc2=3)

Extract the core matrix.

to obtain the core matrix, type:

output$matrizG  
#>           [,1]     [,2]      [,3]      [,4]     [,5]       [,6]
#> [1,] 10.260719 1.847900  3.553432  8.380775 3.021522 -0.5999851
#> [2,] -2.014825 3.989558  3.306571 -1.322206 3.332721 -4.2685767
#> [3,] -1.290695 3.355101 -3.429868  1.325232 3.341179  3.2866310