decay.reciprocal: Decay Functions
In shallot: Random Partition Distribution Indexed by Pairwise Information

Description Usage Arguments Details Author(s) References See Also Examples

View source: R/shallot.R

These functions specify the decay to map distances to attractions.

decay.reciprocal(temperature, distance)

decay.exponential(temperature, distance)

decay.subtraction(temperature, distance, multiplier = 1.01)

## S3 method for class 'shallot.decay'
print(x, ...)

`temperature`	An object of class `shallot.temperature`.
`distance`	An object of class `dist`.
`multiplier`	An scalar greater than `1.0` to ensure that attractions from `decay.subtraction` are finite.
`x`	An object of class `shallot.decay`.
`...`	Currently ignored.

There are currently three choices for decay functions: reciprocal, exponential, and subtraction.

The reciprocal decay maps a distance d to an attraction a as follows: a = 1/d^t, where t is the temperature.

The exponential decay maps a distance d to an attraction a as follows: a = exp(-t*d), where t is the temperature.

The subtract decay maps a distance d to an attraction a as follows: a = (m-d)^t, where t is the temperature and m is the maximum distance in distance multiplied by the supplied multiplier.

David B. Dahl dahl@stat.byu.edu

Dahl, D. B., Day, R., and Tsai, J. (2017), Random Partition Distribution Indexed by Pairwise Information, Journal of the American Statistical Association, 112, 721-732. <DOI:10.1080/01621459.2016.1165103>

dist, temperature, attraction

temp <- temperature(1.0)
distance <- dist(scale(USArrests))
decay1 <- decay.reciprocal(temp,distance)
decay2 <- decay.exponential(temp,distance)
decay3 <- decay.subtraction(temp,distance)