bctsne: Calculate BC t-SNE by orthogonal gradient descent

In daynefiler/bcTSNE: Perform batch-corrected t-SNE

Description

Calculate BC t-SNE by orthogonal gradient descent

Usage

bctsne(X, Z, k = 50, outDim = 2, perplexity = 30, maxIter = 1000)

Arguments

X	numeric matrix, input matrix
Z	numeric matrix, covariate matrix
k	integer of length 1, reduced dimension (number of eigenvectors)
outDim	integer of length 1, the output dimension
perplexity	numeric of length 1, the t-SNE perplexity
maxIter	integer of length 1, the maximum iterations for the BC t-SNE algorithm

Details

X should be preprocessed (e.g. PCA, centered and scaled). Z is the full model matrix, including the intercept.

TODO:

• Add cost function calculation to tsne C code

• Add parameters for max iterations, momentum values, etc.

• Once cost function is calculated, return training data

• Add vector to zero out betas

• Add verbose parameter and print statments in the C code

Value

numeric matrix, t-SNE gradient

daynefiler/bcTSNE documentation built on March 3, 2020, 3:38 p.m.