learn_dbn_structure_pso: Learn a DBN structure with a PSO approach

In dkesada/PSOHO: Particle Swarm Optimization Higher-Order Dynamic Bayesian Network Structure Learning

Description

Given a dataset and the desired Markovian order, this function returns a DBN structure ready to be fitted with the 'dbnR' package. It requires a 'folded' dataset, meaning that all variables have to be in the format 'xxxx_t_y', where 'xxxx' is the name of the variable and 'y' is the time-slice of the variable. This folding can be done manually by shifting the columns and renaming them or automatically via the 'dbnR' package.

Usage

learn_dbn_structure_pso(
  dt,
  size,
  n_inds = 50,
  n_it = 50,
  in_cte = 1,
  gb_cte = 0.5,
  lb_cte = 0.5,
  v_probs = c(10, 65, 25),
  r_probs = c(-0.5, 1.5)
)

Arguments

dt	a data.table with the data of the network to be trained. Previously folded with the 'dbnR' package or other means.
size	Number of timeslices of the DBN. Markovian order 1 equals size 2, and so on.
n_inds	Number of particles used in the algorithm.
n_it	Maximum number of iterations that the algorithm can perform.
in_cte	parameter that varies the effect of the inertia
gb_cte	parameter that varies the effect of the global best
lb_cte	parameter that varies the effect of the local best
v_probs	vector that defines the random velocity initialization probabilities
r_probs	vector that defines the range of random variation of gb_cte and lb_cte

Value

A 'dbn' object with the structure of the best network found

dkesada/PSOHO documentation built on Dec. 7, 2020, 11:35 p.m.