preprocessing {Anthropometry} | R Documentation |
This function allows us to fix the accommodated data before computing archetypes and archetypoids. First, depending on the problem, it is possible to standardize the data or not. Second, it is possible to use the Mahalanobis distance or a depth procedure to select the accommodated subsample of data.
preprocessing(data,stand,percAccomm,mahal=TRUE)
data |
Raw data. It must be a data frame. Each row corresponds to an observation and each column corresponds to an anthropometric variable. All variables are numeric. |
stand |
A logical value. If TRUE (FALSE) the data are (not) standardized. This option will depend on the problem. |
percAccomm |
Percentage of the population to accommodate (value between 0 and 1). When this percentage is equal to 1 all the individuals will be accommodated. |
mahal |
If percAccom is different from 1, then |
In somes cases, the depth procedure has the disadvantage that the desired percentage of accommodation is not under control of the analyst and it could not coincide exactly with percAccomm
.
A list with the following elements if percAccomm
is different from 1:
data: Database after preprocessing, with the 1-percAccomm
percentage of individuals removed.
indivYes: Individuals who belong to data
.
indivNo: Individuals discarded in the accommodation procedure.
A list with the following elements if percAccomm
is equals to 1:
data: Initial database with the same number of observations, which has been standarized depending on the value of stand
.
Irene Epifanio and Guillermo Vinue
Epifanio, I., Vinue, G., and Alemany, S., (2013). Archetypal analysis: contributions for estimating boundary cases in multivariate accommodation problem, Computers & Industrial Engineering 64, 757–765.
Genest, M., Masse, J.-C., and Plante, J.-F., (2012). depth: Depth functions tools for multivariate analysis. R package version 2.0-0.
#As a toy example, only the first 25 individuals are used.
#Variable selection:
variabl_sel <- c(48, 40, 39, 33, 34, 36)
#Changing to inches:
USAFSurvey_inch <- USAFSurvey[1:25, variabl_sel] / (10 * 2.54)
#Data preprocessing:
preproc <- preprocessing(USAFSurvey_inch, TRUE, 0.95, TRUE)
preproc <- preprocessing(USAFSurvey_inch, TRUE, 0.95, FALSE)