mvtsne: Multiview tSNE using an expert opinion pooling on the input...

In dpmccabe/multiview: Multiview Clustering and Dimensionality Reduction Methods

Description

Given a list of of input views and other parameters, mvtsne computes a neighbouring probability matrix for each input view, then finds the optimal set of weights to combine these matrices using a log-linear pool, and applies the pooled probability matrix as input to the standard tSNE procedure, where the probability matrix of the output space is adjusted to the pooled probability matrix using Kullback-Liebler divergence.

Usage

mvtsne(X, k = 2, initial_dims = 30, perplexity = 30, max_iter = 1000,
  min_cost = 0, epoch_callback = NULL, whiten = TRUE, epoch = 100)

Arguments

X	A list of R matrices or "dist" objects, where each matrix/dist is one of the views of the dataset.
k	The desired dimension of the resulting embedding.
initial_dims	Number of dimensions to use in the reduction method.
perplexity	This perplexity parameter is roughly equivalent to the optimal number of neighbours.
max_iter	Maximum number of iterations to perform.
min_cost	The minimum cost value (error) to stop iterations.
epoch_callback	A callback function to be called after each epoch (which is a number of iterations controlled parameter epoch, see next).
whiten	A boolean value indicating if the data matrices should be whitened.
epoch	The number of iterations between update messages.

Value

A list with two elements: embedding with the k-dimensional embedding of the input samples, and weights with the weights associated to each input data view.

Note

All input views must have the same number of samples (rows).

Examples

m1 <- iris[, 1:2]
m2 <- iris[, 3:4]
mvtsne(list(m1, m2), k = 2)

dpmccabe/multiview documentation built on May 5, 2019, 12:30 p.m.