Skip to content

Commit

Permalink
Merge pull request #1522
Browse files Browse the repository at this point in the history 
Manual edits, commenting out some experimental algorithm headers.
  • Loading branch information
@jdramsey
jdramsey authored Feb 14, 2023
2 parents fd96cea + 28a8b65 commit 4e196dd
Show file tree
Hide file tree
Showing 7 changed files with 117 additions and 56 deletions.
145 changes: 103 additions & 42 deletions docs/manual/index.html
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3037,53 +3037,33 @@ <h4>Parameters</h4>
href="#maxPathLength">maxPathLength</a>, <a
href="#completeRuleSetUsed">completeRuleSetUsed</a></p>

<h3>The PAG-Sampling RFCI Algorithm</h3>

<!-- <h3>The RFCI-BSC Algorithm</h3>-->

<!-- <h4>Description</h4>-->

<!-- <div id="rfci-bsc">-->

<!-- <p>RFCI-BSC is a combination of the RFCI [Colombo, 2012]-->
<!-- algorithm and the Bayesian Scoring of Constraints (BSC) method-->
<!-- [Jabbari, 2017] that can generate and probabilistically score-->
<!-- multiple models, outputting the most probable one. This search-->
<!-- algorithm is a hybrid method that derives a Bayesian probability that-->
<!-- the set of independence tests associated with a given causal model-->
<!-- are jointly correct. Using this constraint-based scoring method, we-->
<!-- are able to score multiple causal models, which possibly contain-->
<!-- latent variables, and output the most probable one. See [Jabbari,-->
<!-- 2017].</p>-->

<!-- <p>Currently, this algorithm can only be accessed if experimental-->
<!-- algorithms are activated. One may do so by going to the-->
<!-- File menu, selecting Settings, and then making sure the-->
<!-- 'Experimental' checkbox is checked.</p>-->

<!-- </div>-->
<h4>Description</h4>

<!-- <h4>Input Assumptions</h4>-->
<div id="pag-sampling-rfci">

<p>A modification of the RFCI algorithm which does probabilistic bootstrap
sampling with respect to the RFCI PAG output.For discrete data only.
Parameters are: (a) Number of search probabilistic models, (b) boostrap
ensemble method to use (see bootstrapping), (c) maximimum size of
conditioning set (depth), (d) maximum length of any discriminating path
(a property for RFCI). This must use the probabilistic test, which must
be selected. Parameters for the probabilistic test are (d) independence
cutoff threshold, default 0.5, (e) prior equivalent sample size, and
(f) whether the cutoff in (d) is used in the independence test calculation;
if not, then a coin flip is used (probaility 0.5).
</p>

<!-- <p>The data are discrete only.</p>-->
</div>

<!-- <h4>Output Format</h4>-->
<h4>Input Assumptions</h4>

<!-- <p>A partial ancestral graph (PAG). See Spirtes et al., 2000.</p>-->
<p>A discrete dataset.</p>

<!-- <h4>Parameters</h4>-->

<!-- <p>All of the parameters from RFCI are available for RFCI-BSC.-->
<!-- Additionally:</p>-->
<h4>Output Format</h4>

<!-- <p><a-->
<!-- href="#numRandomizedSearchModels">numRandomizedSearchModels</a>, <a-->
<!-- href="#thresholdNoRandomDataSearch">thresholdNoRandomDataSearch</a>, <a-->
<!-- href="#cutoffDataSearch">cutoffDataSearch</a>, <a-->
<!-- href="#thresholdNoRandomConstrainSearch">thresholdNoRandomConstrainSearch-->
<!-- </a>, <a href="#cutoffConstrainSearch">cutoffConstrainSearch</a>, <a-->
<!-- href="#numBscBootstrapSamples">numBscBootstrapSamples</a>, <a-->
<!-- href="#lowerBound">lowerBound</a>, <a href="#upperBound">upperBound</a>,-->
<!-- <a href="#outputRBD">outputRBD</a></p>-->
<p>A partial ancestral graph (PAG). See Spirtes et al., 2000.</p>

<h3>The GFCI Algorithm</h3>

Expand Down Expand Up @@ -4028,6 +4008,27 @@ <h4>Parameters</h4>

<p><a href="#penaltyDiscount">penaltyDiscount</a></p>

<h3>Kim et al. Scores</h3>

<div id="kim-bic-tests"><p>
This is a set of generalized information criterion (GIC) scores, used as tests,
based on the paper, Kim, Y., Kwon, S., & Choi, H. (2012). Consistent model selection
criteria on high dimensions. The Journal 0of Machine Learning Research,
13(1), 1037-1057. One needs to select which lambda to use in place of the
usual lambda for the linear, Gaussian BIC score. An penalty discount parameter
may also be specified, though this is by default for these scores equal
to 1 (since the lambda choice is essentially picking a penalty discount
for you).
</p>
</div>

<h3>MAG SEM BIC Test</h3>

<div id="mag-sem-bic-test"><p>
This gives a BIC score (used as a test here) for a Mixed Ancestral Graph (MAG).
</p>
</div>

<h3>Probabilistic Test</h3>

<div id="prob-test"><p> The Probabilistic Test applies a Bayesian
Expand Down Expand Up @@ -4354,9 +4355,46 @@ <h3>EBIC Score</h3>
recommended depending on how many variables there are per sample. </p>
</div>

<h4>Parameters</h4>
<h3>Kim et al. Scores</h3>

<div id="kim-scores"><p>
This is a set of generalized information criterion (GIC) scores based on
the paper, Kim, Y., Kwon, S., & Choi, H. (2012). Consistent model selection
criteria on high dimensions. The Journal 0of Machine Learning Research,
13(1), 1037-1057. One needs to select which lambda to use in place of the
usual lambda for the linear, Gaussian BIC score. An penalty discount parameter
may also be specified, though this is by default for these scores equal
to 1 (since the lambda choice is essentially picking a penalty discount
for you).
</p>
</div>

<h3>Poisson Prior Score</h3>

<p><a href="#ebicGamma">ebicGamma</a></p>
<div id="poisson-prior-score"><p>
This is likelihood score attenuated by the log
of the Poisson distribution. It has one parameter, lambda, from the Poisson
distribution, which acts as a structure prior.
</p>
</div>

<h3>Zhang-Shen Bound Score</h3>

<div id="zsbound-score"><p>
Uses Theorem 1 from Zhang, Y., & Shen, X. (2010). Model selection procedure for
high‐dimensional data. Statistical Analysis and Data Mining: The ASA Data
Science Journal, 3(5), 350-358, to make a score that controls false positives.
The is one parameter, the "risk bound", a number between 0 and 1 (a bound on
false positive risk probability).
</p>
</div>

<h3>MAG SEM BIC Score</h3>

<div id="mag-sem-bic-score"><p>
This gives a BIC score for a Mixed Ancestral Graph (MAG).
</p>
</div>

<div id="parameters"><h2>Search Parameters</h2>

Expand Down Expand Up @@ -7240,6 +7278,29 @@ <h3 class="parameter_description" id="precomputeCovariances">precomputeCovarianc
id="penaltyDiscount_value_type">Double</span></li>
</ul>

<h3 class="parameter_description"
id="penaltyDiscountZs">penaltyDiscountZs</h3>
<ul
class="parameter_description_list">
<li>Short Description: <span
id="penaltyDiscountZs_short_desc">Penalty discount (min =
0.0)</span></li>
<li>Long Description: <span
id="penaltyDiscountZs_long_desc"> The parameter c added to a modified
BIC score of the form 2L – c k lambda, where L is the likelihood, k the
number of degrees of freedom, and lambda the choice of GIC lambda. Higher c yield
sparser graphs. </span></li>
<li>Default Value: <span
id="penaltyDiscountZs_default_value">1.0</span></li>
<li>Lower Bound:
<span id="penaltyDiscountZs_lower_bound">0.0</span></li>
<li>Upper
Bound: <span
id="penaltyDiscountZs_upper_bound">1.7976931348623157E308</span></li>
<li>Value Type: <span
id="penaltyDiscountZs_value_type">Double</span></li>
</ul>

<h3 id="zSRiskBound" class="parameter_description">zSRiskBound</h3>
<ul class="parameter_description_list">
<li>Short Description: <span id="zSRiskBound_short_desc">Risk bound</span></li>
Expand Down
10 changes: 5 additions & 5 deletions tetrad-lib/src/main/java/edu/cmu/tetrad/algcomparison/algorithm/oracle/cpdag/BDCE.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,11 +24,11 @@
* @author bryanandrews
* @author josephramsey
*/
@edu.cmu.tetrad.annotation.Algorithm(
name = "BOSS-DCE",
command = "boss-dce",
algoType = AlgType.forbid_latent_common_causes
)
//@edu.cmu.tetrad.annotation.Algorithm(
// name = "BOSS-DCE",
// command = "boss-dce",
// algoType = AlgType.forbid_latent_common_causes
//)
@Bootstrapping
@Experimental
public class BDCE implements Algorithm, UsesScoreWrapper {
Expand Down
10 changes: 5 additions & 5 deletions tetrad-lib/src/main/java/edu/cmu/tetrad/algcomparison/algorithm/oracle/cpdag/BOSSDC.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,11 +25,11 @@
* @author bryanandrews
* @author josephramsey
*/
@edu.cmu.tetrad.annotation.Algorithm(
name = "BOSS-DC",
command = "boss-dc",
algoType = AlgType.forbid_latent_common_causes
)
//@edu.cmu.tetrad.annotation.Algorithm(
// name = "BOSS-DC",
// command = "boss-dc",
// algoType = AlgType.forbid_latent_common_causes
//)
@Bootstrapping
@Experimental
public class BOSSDC implements Algorithm, UsesScoreWrapper {
Expand Down
1 change: 0 additions & 1 deletion tetrad-lib/src/main/java/edu/cmu/tetrad/algcomparison/algorithm/oracle/pag/GRASP_FCI.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,6 @@
algoType = AlgType.allow_latent_common_causes
)
@Bootstrapping
@Experimental
public class GRASP_FCI implements Algorithm, UsesScoreWrapper, TakesIndependenceWrapper, HasKnowledge {

static final long serialVersionUID = 23L;
Expand Down
2 changes: 1 addition & 1 deletion tetrad-lib/src/main/java/edu/cmu/tetrad/algcomparison/independence/KimEtAlScoreTests.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@ public DataType getDataType() {
public List<String> getParameters() {
List<String> params = new ArrayList<>();
params.add(Params.SEM_GIC_RULE);
params.add(Params.PENALTY_DISCOUNT);
params.add(Params.PENALTY_DISCOUNT_ZS);
return params;
}
}
4 changes: 2 additions & 2 deletions tetrad-lib/src/main/java/edu/cmu/tetrad/algcomparison/score/KimEtAlScores.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@
* @author jdramsey
*/
@edu.cmu.tetrad.annotation.Score(
name = "Kim et al. Score",
name = "Kim et al. Scores",
command = "kim-scores",
dataType = {DataType.Continuous, DataType.Covariance}
)
Expand Down Expand Up @@ -89,7 +89,7 @@ public DataType getDataType() {
public List<String> getParameters() {
List<String> parameters = new ArrayList<>();
parameters.add(Params.SEM_GIC_RULE);
parameters.add(Params.PENALTY_DISCOUNT);
parameters.add(Params.PENALTY_DISCOUNT_ZS);
return parameters;
}

Expand Down
1 change: 1 addition & 0 deletions tetrad-lib/src/main/java/edu/cmu/tetrad/util/Params.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -147,6 +147,7 @@ public final class Params {
public static final String OUTPUT_RBD = "outputRBD";
public static final String PARALLELIZED = "parallelized";
public static final String PENALTY_DISCOUNT = "penaltyDiscount";
public static final String PENALTY_DISCOUNT_ZS = "penaltyDiscountZs";
public static final String EBIC_GAMMA = "ebicGamma";
public static final String PERCENT_DISCRETE = "percentDiscrete";
public static final String PERCENT_RESAMPLE_SIZE = "percentResampleSize";
Expand Down

0 comments on commit 4e196dd

Please sign in to comment.