CV_delta: Cross validation to select delta
In bingx1990/LOVE: Overlapping clustering based on latent factor models

Description Usage Arguments Value

View source: R/CV.R

Cross validation for choosing the tuning parameter δ. For each value of deltaGrids, first split the data into two parts and calculate I, A_I and Cov(Z). Then calculate the fit A_I Cov(Z) A_I' to find the value which minimizes the loss criterion

||Σ - A_I Cov(Z) A_I'||_{F-off}/(|I|(|I|-1))

.

1	CV_delta(X, deltaGrids, diagonal, se_est, merge)

`X`	A n by p data matrix.
`deltaGrids`	A vector of numerical constants.
`diagonal`	Logical. If TRUE, the covariance matrix of Z is diagonal; else FALSE.
`se_est`	The vector of the standard deviations of p features.
`merge`	Logical. If TRUE, take the union of all candidate pure variables; otherwise, take the intersection.