qdtsne Namespace Reference

Quick and dirty t-SNE. More...

Classes
struct	Options
	Options for initialize(). More...
class	Status
	Status of the t-SNE iterations. More...

Typedefs
template<typename Index_ , typename Float_ >
using	NeighborList = knncolle::NeighborList<Index_, Float_>
	Lists of neighbors for each observation.

Functions
template<int num_dim_, typename Index_ , typename Float_ >
Status< num_dim_, Index_, Float_ >	initialize (NeighborList< Index_, Float_ > neighbors, const Options &options)
template<int num_dim_, typename Dim_ , typename Index_ , typename Float_ >
Status< num_dim_, Index_, Float_ >	initialize (const knncolle::Prebuilt< Dim_, Index_, Float_ > &prebuilt, const Options &options)
template<int num_dim_, typename Dim_ , typename Index_ , typename Float_ >
Status< num_dim_, Index_, Float_ >	initialize (Dim_ data_dim, Index_ num_points, const Float_ *data, const knncolle::Builder< knncolle::SimpleMatrix< Dim_, Index_, Float_ >, Float_ > &builder, const Options &options)
int	perplexity_to_k (double perplexity)
template<int num_dim_, typename Float_ = double>
void	initialize_random (Float_ *Y, size_t num_points, int seed=42)
template<int num_dim_, typename Float_ = double>
std::vector< Float_ >	initialize_random (size_t num_points, int seed=42)
template<typename Task_ , class Run_ >
void	parallelize (int num_workers, Task_ num_tasks, Run_ run_task_range)

Detailed Description

Quick and dirty t-SNE.

Typedef Documentation

NeighborList

template<typename Index_ , typename Float_ >

using qdtsne::NeighborList = knncolle::NeighborList<Index_, Float_>

Lists of neighbors for each observation.

This is a convenient alias for the knncolle::NeighborList class. Each inner vector corresponds to an observation and contains the list of nearest neighbors for that observation, sorted by increasing distance. Neighbors for each observation should be unique - there should be no more than one occurrence of each index in each inner vector. Also, the inner vector for observation i should not contain any Neighbor with index i.

Template Parameters

Index_	Integer type to use for the indices.
Float_	Floating-point type to use for the calculations.

Function Documentation

initialize() [1/3]

template<int num_dim_, typename Dim_ , typename Index_ , typename Float_ >

Status< num_dim_, Index_, Float_ > qdtsne::initialize	(	const knncolle::Prebuilt< Dim_, Index_, Float_ > &	prebuilt,
		const Options &	options )

Overload that accepts a neighbor search index and computes the nearest neighbors for each observation, before proceeding with the initialization of the t-SNE algorithm.

Template Parameters

num_dim_	Number of dimensions of the final embedding.
Dim_	Integer type for the dataset dimensions.
Index_	Integer type for the neighbor indices.
Float_	Floating-point type to use for the calculations.

Parameters

prebuilt	A knncolle::Prebuilt instance containing a neighbor search index built on the dataset of interest.
options	Further options.

Returns

A Status object representing an initial state of the t-SNE algorithm.

initialize() [2/3]

template<int num_dim_, typename Dim_ , typename Index_ , typename Float_ >

Status< num_dim_, Index_, Float_ > qdtsne::initialize	(	Dim_	data_dim,
		Index_	num_points,
		const Float_ *	data,
		const knncolle::Builder< knncolle::SimpleMatrix< Dim_, Index_, Float_ >, Float_ > &	builder,
		const Options &	options )

Overload that accepts a column-major matrix of coordinates and computes the nearest neighbors for each observation, before proceeding with the initialization of the t-SNE algorithm.

Template Parameters

num_dim_	Number of dimensions of the final embedding.
Dim_	Integer type for the dataset dimensions.
Index_	Integer type for the neighbor indices.
Float_	Floating-point type to use for the calculations.

Parameters

	data_dim	Number of rows of the matrix at data, corresponding to the dimensions of the input dataset.
	num_points	Number of columns of the matrix at data, corresponding to the points of the input dataset.
[in]	data	Pointer to an array containing a column-major matrix with data_dim rows and num_points columns.
	builder	A knncolle::Builder instance specifying the nearest-neighbor algorithm to use.
	options	Further options.

Returns

A Status object representing an initial state of the t-SNE algorithm.

initialize() [3/3]

template<int num_dim_, typename Index_ , typename Float_ >

Status< num_dim_, Index_, Float_ > qdtsne::initialize	(	NeighborList< Index_, Float_ >	neighbors,
		const Options &	options )

Initialize the data structures for t-SNE algorithm, given the nearest neighbors of each observation.

Template Parameters

num_dim_	Number of dimensions of the final embedding.
Index_	Integer type for the neighbor indices.
Float_	Floating-point type to use for the calculations.

Parameters

neighbors	List of indices and distances to nearest neighbors for each observation. Each observation should have the same number of neighbors, sorted by increasing distance, which should not include itself.
options	Further options. If Options::infer_perplexity = true, the perplexity is determined from neighbors and the value in Options::perplexity is ignored.

Returns

A Status object representing an initial state of the t-SNE algorithm.

initialize_random() [1/2]

template<int num_dim_, typename Float_ = double>

void qdtsne::initialize_random	(	Float_ *	Y,
		size_t	num_points,
		int	seed = 42 )

Initializes the starting locations of each observation in the embedding. We do so using our own implementation of the Box-Muller transform, to avoid problems with differences in the distribution functions across C++ standard library implementations.

Template Parameters

num_dim_	Number of embedding dimensions.
Float_	Floating-point type to use for the calculations.

Parameters

[out]	Y	Pointer to a 2D array with number of rows and columns equal to num_dim and num_points, respectively. On output, Y is filled with random draws from a standard normal distribution.
	num_points	Number of points in the embedding.
	seed	Seed for the random number generator.

initialize_random() [2/2]

template<int num_dim_, typename Float_ = double>

std::vector< Float_ > qdtsne::initialize_random	(	size_t	num_points,
		int	seed = 42 )

Creates the initial locations of each observation in the embedding.

Template Parameters

num_dim_	Number of embedding dimensions.
Float_	Floating-point type to use for the calculations.

Parameters

num_points	Number of observations.
seed	Seed for the random number generator.

Returns

A vector of length num_points * num_dim_ containing random draws from a standard normal distribution.

parallelize()

template<typename Task_ , class Run_ >

void qdtsne::parallelize	(	int	num_workers,
		Task_	num_tasks,
		Run_	run_task_range )

Template Parameters

Task_	Integer type for the number of tasks.
Run_	Function to execute a range of tasks.

Parameters

num_workers	Number of workers.
num_tasks	Number of tasks.
run_task_range	Function to iterate over a range of tasks within a worker.

By default, this is an alias to subpar::parallelize_range(). However, if the QDTSNE_CUSTOM_PARALLEL function-like macro is defined, it is called instead. Any user-defined macro should accept the same arguments as subpar::parallelize_range().

perplexity_to_k()

int qdtsne::perplexity_to_k ( double perplexity )

inline

Determines the appropriate number of neighbors, given a perplexity value. Useful when the neighbor search is conducted outside of initialize().

Parameters

perplexity

Perplexity to use in the t-SNE algorithm.

Returns

Number of nearest neighbors to find.