slide: Fits SLIDE model to the multi-view data X

In irinagain/SLIDE: Structural Learning and Integrative Decomposition of Multi-View Data

Description

Fits SLIDE model to the multi-vew data X

Usage

slide(
  X,
  pvec,
  n_lambda = 50,
  lambda_min = 0.01,
  n_fold = 3,
  p_fold = 3,
  center = T,
  k_max = 5000,
  eps = 1e-06,
  ratio_max = NULL
)

Arguments

X	A n x p concatenated data matrix of views X_1,...,X_d.
pvec	A vector of values p_1,....,p_d corresponding to the number of measurements within each view.
n_lambda	A length of tuning parameter sequence used for generation of candidate structures. The default value is 50.
lambda_min	A minimal value for tuning parameter. The default value is 0.01.
n_fold	A number of folds for rows in BCV procedure, the default value is 3.
p_fold	A number of folds for columns in BCV procedure, the default value is 3.
center	A logical indicator of whether the data should be centered, the default value is TRUE.
k_max	A maximal number of allowed iterations for SLIDE model fitting, the default value is 1000.
eps	A convergence tolerance criterion, the default value is 1e-6.
ratio_max	A maximal allowable rank of the binary structure as a ratio of maximal rank of X, the default value is set based on bi-cross-validation folds; with 3 folds corresponding to ratio_max = 2/3 and 2 folds corresponding to ratio_max = 1/2.

Details

The function automatically column-centers and standardizes X so that the Frobenius norm within each view is equal to one (using function standardizeX). The function then performs 3 steps of SLIDE workflow:

1. Create a candidate list of distinct binary structures S_1,...,S_m (using function create_structure_list).

2. Select one structure S from the list using the bi-cross-validation (using function slideBCV).

3. Fit SLIDE model with selected structure S (using function slide_givenS).

Value

A list with the elements

out_s	Output from standardizedX applied to X.
structure_list	A list of considered candidate structures.
S	A selected binary structure for the SLIDE model.
model	Slide model fitted on selected S, output from slide_givenS.

Examples

n = 25

# Two matched datasets
p1 = 10
p2 = 10
data = generateModel1(n = n, pvec = c(p1, p2))
out_slide = slide(X = data$X, pvec = c(p1,p2))
out_slide$S

# Three matched datasets
p3 = 15
data = generateModel2(n = n, pvec = c(p1, p2, p3))
out_slide = slide(X = data$X, pvec = c(p1, p2, p3))
out_slide$S

irinagain/SLIDE documentation built on Aug. 14, 2021, 2:56 p.m.