ndr function

Genearlized Network-based Dimensionality Reduction and Analysis (GNDA)

Genearlized Network-based Dimensionality Reduction and Analysis (GNDA)

The main function of Generalized Network-based Dimensionality Reduction and Analysis (GNDA).

ndr(r,covar=FALSE,cor_method=1,cor_type=1,min_R=0,min_comm=2,Gamma=1,null_model_type=4,
mod_mode=6,min_evalue=0,min_communality=0,com_communalities=0,use_rotation=FALSE,
rotation="oblimin",weight=NULL,seed=NULL)

Arguments

  • r: A numeric data frame
  • covar: If this value is FALSE (default), it finds the correlation matrix from the raw data. If this value is TRUE, it uses the matrix r as a correlation/similarity matrix.
  • cor_method: Correlation method (optional). '1' Pearson's correlation (default), '2' Spearman's correlation, '3' Kendall's correlation, '4' Distance correlation
  • cor_type: Correlation type (optional). '1' Bivariate correlation (default), '2' partial correlation, '3' semi-partial correlation
  • min_R: Minimal square correlation between indicators (default: 0).
  • min_comm: Minimal number of indicators per community (default: 2).
  • Gamma: Gamma parameter in multiresolution null modell (default: 1).
  • null_model_type: '1' Differential Newmann-Grivan's null model, '2' The null model is the mean of square correlations between indicators, '3' The null model is the specified minimal square correlation, '4' Newmann-Grivan's modell (default)
  • mod_mode: Community-based modularity calculation mode: '1' Louvain modularity, '2' Fast-greedy modularity, '3' Leading Eigen modularity, '4' Infomap modularity, '5' Walktrap modularity, '6' Leiden modularity (default)
  • min_evalue: Minimal eigenvector centrality value (default: 0)
  • min_communality: Minimal communality value of indicators (default: 0)
  • com_communalities: Minimal common communalities (default: 0)
  • use_rotation: FALSE no rotation (default), TRUE the rotation is used.
  • rotation: "none", "varimax", "quartimax", "promax", "oblimin", "simplimax", and "cluster" are possible rotations/transformations of the solution. "oblimin" is the default, if use_rotation is TRUE.
  • weight: The weights of columns. The defalt is NULL (no weights).
  • seed: default seed value (default=NULL, no seed)

Details

NDA both works on low and high simple size datasets. If min_evalue=min_communality=com_communalities=0 than there is no feature selection.

Returns

  • communality: Communality estimates for each item. These are merely the sum of squared factor loadings for that item. It can be interpreted in correlation matrices.

  • loadings: A standard loading matrix of class “loadings".

  • uniqueness: Uniqueness values of indicators.

  • factors: Number of found factors.

  • EVCs: The list eigenvector centrality value of indicators.

  • membership: The membership value of indicators.

  • weight: The weight of indicators.

  • scores: Estimates of the factor scores are reported (if covar=FALSE).

  • centers: Colum mean of unstandardized score values.

  • n.obs: Number of observations specified or found.

  • use_rotation: FALSE no rotation (default), TRUE the rotation is used.

  • rotation: "none", "varimax", "quartimax", "promax", "oblimin", "simplimax", and "cluster" are possible rotations/transformations of the solution. "oblimin" is the default, if use_rotation is TRUE.

  • fn: Factor name: NDA

  • seed: applied seed value (default=NULL, no seed)

  • Call: Callback function

References

Kosztyan, Z. T., Kurbucz, M. T., & Katona, A. I. (2022). Network-based dimensionality reduction of high-dimensional, low-sample-size datasets. Knowledge-Based Systems, 109180. doi:10.1016/j.knosys.2022.109180

Author(s)

Zsolt T. Kosztyan*, Marcell T. Kurbucz, Attila I. Katona

e-mail*: kosztyan.zsolt@gtk.uni-pannon.hu

See Also

plot, biplot, summary.

Examples

# Dimension reduction without using any hyperparameters

data(swiss)
df<-swiss
p<-ndr(df)
summary(p)
plot(p)
biplot(p)

# Dimension reduction with using hyperparameters
# min_R=0.1 # The mininal square correlation must be grater than 0.1

p<-ndr(df,min_R = 0.1)
summary(p)
plot(p)

# min_evalue=0.1 # Minimal evector centalities must be greater than 0.1

p<-ndr(df,min_evalue = 0.1)
summary(p)
plot(p)

# minimal and common communality value must be greater than 0.25

p<-ndr(df,min_communality = 0.25,
 com_communalities = 0.25)

# Print factor matrix
cor(p$scores)
plot(p)

# Use factor rotation

p<-ndr(df,min_communality = 0.25,
 com_communalities = 0.25,use_rotation=TRUE)

# Print factor matrix
cor(p$scores)
biplot(p)

# Data reduction - clustering
# Distance is Euclidean's distance
# covar=TRUE means only the distance matrix is considered.

q<-ndr(1-normalize(as.matrix(dist(df))),covar=TRUE)
summary(q)
plot(q)
nda packageRead PDF manual