In [14]:

    

library("bnstruct")
library("dagR")


    

In [2]:

    

count <- function(x) {
    return(length(unique(x)))
}

quantize <- function(x) {
    return(x - min(x) + 1)
}


    

In [18]:

    

# Given model:
#
#    A -> B
#    A -> C
#    B -> D
#    C -> D
#

SIZE <- 1000

A <- round(rnorm(SIZE, 10, 2))
B <- round(0.4 * A + rnorm(SIZE, 10, 1))
C <- round(0.8 * A + rnorm(SIZE, 10, 1.5))
D <- round(0.4 * B + 0.7 * C + rnorm(SIZE, 10, 1))

data = cbind(A, B, C, D)
data = apply(data, 2, quantize)

dataset <- BNDataset(data = as.matrix(data),
                     discreteness = rep('D', 4),
                     variables = c("A", "B", "C", "D"),
                     node.sizes = apply(data, 2, count))

dataset_bs <- bootstrap(dataset, num.boots = 1000)


    

    




bnstruct :: Generating bootstrap samples ...
bnstruct :: Bootstrap samples generated.

In [19]:

    

net <- learn.network(dataset, scoring.func = "BDeu", algo="SEM")
dag(net)
plot(net)


    

    




bnstruct :: learning the structure using SEM ...
... bnstruct :: starting EM algorithm ...
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... bnstruct :: EM algorithm completed.
... bnstruct :: learning the structure using MMHC ...
... bnstruct :: learning using MMHC completed.
... bnstruct :: learning network parameters ... 
... bnstruct :: parameter learning done.
... bnstruct :: starting EM algorithm ...
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... bnstruct :: EM algorithm completed.
... bnstruct :: learning the structure using MMHC ...
... bnstruct :: learning using MMHC completed.
... bnstruct :: learning network parameters ... 
... bnstruct :: parameter learning done.
bnstruct :: learning using SEM completed.
bnstruct :: learning network parameters ... 
bnstruct :: parameter learning done.






    







	
0
1
0
0
	
0
0
0
0
	
1
0
0
1
	
0
0
0
0









    

In [25]:

    

net <- learn.network(dataset_bs, bootstrap = TRUE)
wpdag(net)
#plot(net, plot.wpdag=T)
plot(net)


    

    




bnstruct :: learning the structure using MMHC ...
bnstruct :: learning using MMHC completed.






    







	
  0 
 439
924 
   0
	
978 
   0
  0 
1000
	
999 
   0
  0 
   1
	
  0 
1000
  3 
   0









    

In [22]:

    

# Given model:
#
#    A -> B
#    A -> C
#    B -> D
#    C -> D
#

SIZE <- 1000

A <- round(rnorm(SIZE, 10, 2))
B <- round(0.4 * A + rnorm(SIZE, 10, 1))
C <- round(0.8 * A + rnorm(SIZE, 10, 1.5))
D <- round(0.4 * B + 0.2 * C + rnorm(SIZE, 10, 1))

data = cbind(A, B, C, D)
data = apply(data, 2, quantize)

dataset <- BNDataset(data = as.matrix(data),
                     discreteness = rep('D', 4),
                     variables = c("A", "B", "C", "D"),
                     node.sizes = apply(data, 2, count))

dataset_bs <- bootstrap(dataset, num.boots = 1000)


    

    




bnstruct :: Generating bootstrap samples ...
bnstruct :: Bootstrap samples generated.

In [23]:

    

net <- learn.network(dataset, scoring.func = "BDeu", algo="SEM")
dag(net)
plot(net)


    

    




bnstruct :: learning the structure using SEM ...
... bnstruct :: starting EM algorithm ...
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... bnstruct :: EM algorithm completed.
... bnstruct :: learning the structure using MMHC ...
... bnstruct :: learning using MMHC completed.
... bnstruct :: learning network parameters ... 
... bnstruct :: parameter learning done.
... bnstruct :: starting EM algorithm ...
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... ... bnstruct :: learning network parameters ... 
... ... bnstruct :: parameter learning done.
... bnstruct :: EM algorithm completed.
... bnstruct :: learning the structure using MMHC ...
... bnstruct :: learning using MMHC completed.
... bnstruct :: learning network parameters ... 
... bnstruct :: parameter learning done.
bnstruct :: learning using SEM completed.
bnstruct :: learning network parameters ... 
bnstruct :: parameter learning done.






    







	
0
0
1
0
	
1
0
0
1
	
0
0
0
0
	
0
0
0
0









    

In [ ]: