| Title: | Bayesian Network Learning Improved Project |
| Version: | 1.1 |
| Description: | Allows the user to learn Bayesian networks from datasets containing thousands of variables. It focuses on score-based learning, mainly the 'BIC' and the 'BDeu' score functions. It provides state-of-the-art algorithms for the following tasks: (1) parent set identification - Mauro Scanagatta (2015) http://papers.nips.cc/paper/5803-learning-bayesian-networks-with-thousands-of-variables; (2) general structure optimization - Mauro Scanagatta (2018) <doi:10.1007/s10994-018-5701-9>, Mauro Scanagatta (2018) http://proceedings.mlr.press/v73/scanagatta17a.html; (3) bounded treewidth structure optimization - Mauro Scanagatta (2016) http://papers.nips.cc/paper/6232-learning-treewidth-bounded-bayesian-networks-with-thousands-of-variables; (4) structure learning on incomplete data sets - Mauro Scanagatta (2018) <doi:10.1016/j.ijar.2018.02.004>. Distributed under the LGPL-3 by IDSIA. |
| Depends: | R (≥ 3.0.0) |
| Imports: | foreign, bnlearn (≥ 4.0) |
| SystemRequirements: | Java (>= 1.5) |
| License: | LGPL-3 |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 6.1.1 |
| NeedsCompilation: | no |
| Packaged: | 2019-02-19 13:03:35 UTC; loskana |
| Author: | Mauro Scanagatta [aut, cre] |
| Maintainer: | Mauro Scanagatta <mauro@idsia.ch> |
| Repository: | CRAN |
| Date/Publication: | 2019-02-27 19:20:21 UTC |
Bayesian Learning Package - Main function.
Description
Used by most of the functions in the r.blip binding, provides access to the included jar file.
Usage
blip(args)
Arguments
args |
Vector of arguments to be passed to the jar |
Details
The arguments vector is formatted in a system call to the included jar file. Should not be called directly by the user, unless you know exactly what you are doing. In that case, call directly the blip jar.
Learns a BN
Description
Fully learns a Bayesian networks.
Usage
blip.learn(dat, scorer.method = "is", solver.method = "winasobs",
indeg = 6, time = 3600, allocated = 80, scorefunction = "bic",
alpha = 1, cores = 1, verbose = 0)
Arguments
dat |
dataframe from which to learn the parent sets.(required) |
scorer.method |
Method to be used for scoring the parent sets. Possible values: "is" (independence selection), "sq" (sequential selection). (default: is) |
solver.method |
Method to be used for structure exploration. Possible values: "winasobs", "winobs", "asobs", "obs". (default: winasobs) |
indeg |
Maximum number of parents (default: 6) |
time |
Execution time (default: 3600) |
allocated |
Percentage of the total execution time dedicated to parent set exploration (default: 80) |
scorefunction |
Chosen score function. Possible choices: BIC, BDeu (default: bic) |
alpha |
(if BDeu is chosen) equivalent sample size parameter (default: 1.0) |
cores |
Number of machine cores to use. If 0, all are used. (default: 1) |
verbose |
Verbose level (default: 0) |
Details
The input data is required to be complete and discrete. Accordingly missing values in the input data.frame will be ignored, and all numeric values will be converted to integers.
Value
The learned Bayesian network in the bnlearn format.
Examples
bn <- blip.learn(child, time=3)
Learns a BN with a treewidth bound
Description
Fully learns a Bayesian networks with a treewidth bound.
Usage
blip.learn.tw(dat, scorer.method = "is", solver.method = "kmax",
treewidth = 5, time = 3600, allocated = 80,
scorefunction = "bic", alpha = 1, cores = 1, verbose = 0)
Arguments
dat |
dataframe from which to learn the parent sets.(required) |
scorer.method |
Method to be used for scoring the parent sets. Possible values: "is" (independence selection), "sq" (sequential selection). (default: is) |
solver.method |
Method to be used for bounded-treewidth structure exploration. Possible values: "kmax", "kg", "ka". (default: kmax) |
treewidth |
Maximum treewidth (default: 4) |
time |
Execution time (default: 3600) |
allocated |
Percentage of the total execution time dedicated to parent set exploration (default: 80) |
scorefunction |
Chosen score function. Possible choices: BIC, BDeu (default: bic) |
alpha |
(if BDeu is chosen) equivalent sample size parameter (default: 1.0) |
cores |
Number of machine cores to use. If 0, all are used. (default: 1) |
verbose |
Verbose level (default: 0) |
Details
The input data is required to be complete and discrete. Accordingly missing values in the input data.frame will be ignored, and all numeric values will be converted to integers.
Value
The learned Bayesian network in the bnlearn format.
Examples
bn <- blip.learn.tw(child, treewidth=4, time=3)
Parent set exploration
Description
Generates the cache of parent sets from a given data source
Usage
blip.scorer(dat, method = "is", indeg = 6, time = 3600,
scorefunction = "bic", alpha = 1, cores = 1, verbose = 0)
Arguments
dat |
dataframe from which to learn the parent sets.(required) |
method |
Method to be used. Possible values: "is" (independence selection), "sq" (sequential selection). (default: is) |
indeg |
Maximum number of parents (default: 6) |
time |
Maximum Execution time (default: 3600) |
scorefunction |
Chosen score function. Possible choices: BIC, BDeu (default: bic) |
alpha |
(if BDeu is chosen) equivalent sample size parameter (default: 1.0) |
cores |
Number of machine cores to use. If 0, all are used. (default: 1) |
verbose |
Verbose level (default: 0) |
Details
Usually the first step in the learning of a Bayesian network.
The input data is required to be complete and discrete. Accordingly missing values in the input data.frame will be ignored, and all numeric values will be converted to integers.
Value
Cache of parent sets
Examples
jkl <- blip.scorer(child, time=3)
Structure Optimization
Description
Find an optimal structure from the cache of parent sets
Usage
blip.solver(jkl, method = "winasobs", time = 3600, cores = 1,
verbose = 0)
Arguments
jkl |
cache of pre-computed parent sets.(required) |
method |
Method to be used. Possible values: "winasobs", "winobs", "asobs", "obs". (default: winasobs) |
time |
Maximum Execution time (default: 3600) |
cores |
Number of machine cores to use. If 0, all are used. (default: 1) |
verbose |
Verbose level (default: 0) |
Details
The input data is required to be complete and discrete. Accordingly missing values in the input data.frame will be ignored, and all numeric values will be converted to integers.
Value
Structure
Examples
bn <- blip.solver(child.jkl, time=3)
Structure Optimization - treewidth bound
Description
Find an optimal structure from the cache of parent sets
Usage
blip.solver.tw(jkl, method = "kmax", treewidth = 4, time = 3600,
cores = 1, verbose = 0)
Arguments
jkl |
cache of pre-computed parent sets.(required) |
method |
Method to be used. Possible values: "kmax", "kg", "ka". (default: kmax) |
treewidth |
Maximum treewidth (default: 4) |
time |
Maximum Execution time (default: 3600) |
cores |
Number of machine cores to use. If 0, all are used. (default: 1) |
verbose |
Verbose level (default: 0) |
Details
The input data is required to be complete and discrete. Accordingly missing values in the input data.frame will be ignored, and all numeric values will be converted to integers.
Value
Structure
Examples
bn <- blip.solver.tw(child.jkl, time=3)
Child dataset
Description
Dataset generated from the famous "child" network. Provided as an example of input data for learning a Bayesian network.
Usage
data("child")Format
The format is: chr "child"
Details
Space separeted, integer values for each variable.
Source
http://www.bnlearn.com/bnrepository/discrete-medium.html#child
Examples
data(child)
Parent set cache for the child dataset
Description
Parent set cache, taken from the "child" dataset.
Jkl reader
Description
Read a Jkl file (parent sets cache)
Usage
read.jkl(path, names)
Arguments
path |
Path of the file to load |
names |
List of variable names |
Value
the cache of parent sets
Structure reader
Description
Reads a str file (BN structure)
Usage
read.str(path, names)
Arguments
path |
Path of the file to load |
names |
List of variable names |
Value
the BN structure
Jkl writer (with names)
Description
Write a Jkl file (parent sets cache)
Usage
write.jkl(path, jkl)
Arguments
path |
Path of the file to write |
jkl |
parent sets cache to write |