# DimensionReduction: 'DimensionReduction' In NumbersInternational/flipDimensionReduction: Algorithms for dimension reduction

 DimensionReduction R Documentation

## `DimensionReduction`

### Description

Produces as 2-dimensional embedding. Takes either a `data.frame` of variables and optionally another `vector` to be used as a grouping variable, or takes a distance `matrix`.

### Usage

``````DimensionReduction(
algorithm,
data = NULL,
data.groups = NULL,
table = NULL,
raw.table = FALSE,
subset = NULL,
perplexity = 10,
binary = TRUE,
normalization = FALSE,
seed = 1066,
print.type = "2d",
use.combined.scatter = FALSE,
...
)
``````

### Arguments

 `algorithm` Valid options are `"t-SNE"`, `"MDS - Metric"`, `"MDS - Non-metric"`, or `"PCA"`. Note that `'PCA'` does not accept a distance matrix as input. `data` A `data.frame` with cases by row and variables by column. `data.groups` A `vector` to be used as a grouping variable for the embedded `data`. `table` A symmetrical distance `matrix` or an object of class `dist` or `DistanceMatrix`. `raw.table` If `TRUE`, `ParseUserEnteredTable` is called on `table` to create a numeric matrix from text. `subset` A logical vector which describes the subset of `data` to be analyzed. Not used for `table` input. `perplexity` The perplexity coefficient which defines the extent of the locality of the dimension reduction. Used only when `algorithm` is `t-SNE`. `binary` If `TRUE`, unordered factors are converted to dummy variables. Otherwise, their value attributes/levels are used to coerce each factor to a single numeric variable. See the Details of `AsNumeric`. `normalization` If `data` is supplied, whether to standardize the data so each variable has a mean of 0 and standard deviation of 1. `seed` Random seed. Used only when `algorithm` is `"t-SNE"`. `print.type` Specifies output produced if algorithm is PCA. `use.combined.scatter` Draw scatterplots using rhtmlCombinedScatter. `...` Other parameters passed to PrincipalComponentsAnalysis.

### Details

For `data` input, all algorithms apart from `PCA` remove duplicated data and any case with `NA` is ignored by all algorithms.

NumbersInternational/flipDimensionReduction documentation built on Aug. 20, 2024, 2:32 a.m.