Assign a distance measure

Assigns a distance measure to the MapperRef instance to use in the clustering algorithm.

Usage

 $use_distance_measure(measure, ...) 

Arguments

measure

The distance measure to use (string).

...

Extra parameters passed to the distance function. See details.

Details

Unless the clustering_algorithm has been replaced by the user, by default, Mapper requires a notion of distance between objects to be defined in creating the vertices of the construction.

The distance function is determined based on the supplied measure. measure must be one of:

["euclidean", "maximum", "manhattan", "canberra", "binary", "minkowski"]

or, if the proxy and parallelDist packages are installed, any name in proxy::pr_DB$get_entry_names().

Additional parameters passed via ... are passed to either dist (or parallelDist if installed).

Value

The mapper instance, with the measure field assigned.

See also

parDist pr_DB

Examples

data(noisy_circle)
m <- MapperRef$new(noisy_circle)
m$use_filter(noisy_circle[,1])
m$use_cover("fixed interval", number_intervals = 5, percent_overlap = 25)

## Constructs clusters with euclidean metric (default)
m$use_distance_measure("euclidean")
m$construct_pullback()

## Constructs clusters with p-norm (p = 1)
m$use_distance_measure("Minkowski", p = 1L)
m$construct_pullback()
# NOT RUN {
## To see list of available measures, use:
proxy::pr_DB$get_entry_names()
# }