particles: Initialize particles to perform inference in a Gaussian...

In gmgm: Gaussian Mixture Graphical Model Learning and Inference

Initialize particles to perform inference in a Gaussian mixture graphical model

Description

This function initializes particles to perform (approximate) inference in a Gaussian mixture graphical model. Particles consist in weighted sample sequences propagated forward in time by sampling the model and aggregated to obtain the inferred values (Koller and Friedman, 2009).

Usage

particles(seq = NULL, col_weight = "weight", n_part = 1000)

Arguments

seq	A data frame containing the observation sequences for which particles are initialized. If NULL (the default), the initialization is performed for a single sequence.
col_weight	A character string corresponding to the column name of the resulting data frame that describes the particle weight.
n_part	A positive integer corresponding to the number of particles initialized for each observation sequence.

Value

A data frame (tibble) containing the initial particles.

References

Koller, D. and Friedman, N. (2009). Probabilistic Graphical Models: Principles and Techniques. The MIT Press.

See Also

aggregation, propagation

Examples

data(data_air)
part <- particles(data.frame(DATE = unique(data_air$DATE)))

gmgm documentation built on Sept. 9, 2022, 1:07 a.m.