Embedding

Description

The embedding function extracts a low-dimensional representation from the results of a singular value decomposition (SVD). This function is useful for dimensionality reduction, creating a compact embedding that preserves significant data characteristics.

embedding<-function(fit,r){
  embed <- fit$u[,1:r]%*%diag(sqrt(fit$d[1:r]))
  embed
}

Arguments

• fit: An object containing the SVD results, with components u (left singular vectors) and d (singular values).
• r: The target dimensionality for the embedding (number of components to retain).

Output

• Returns a matrix representing the embedding of the data, obtained by multiplying the first r columns of u with the diagonal matrix of the square root of the first r singular values.

See also

See BONMI_package for code example.