ChoR.Budget {ChoR}

Call to the budget chordalysis algorithm.

Description

Searches a statistically significant decomposable model to explain a dataset using Prioritized Chordalysis.

Usage

ChoR.Budget(x, pValueThreshold = 0.05, budgetShare = 0.01, card = NULL)

Arguments

x

A dataframe with categorical data; column names are the name of the attributes.

pValueThreshold

A double value, minimum p-value for statistical consistency (commonly 0.05)

budgetShare

A double value, share of the statistical budget to consume at each step (>0 and <=1; 0.01 seems like a reasonable value for most datasets)

card

A vector containing the cardinality of the attributes (position wise).

Details

Call the Budget chordalysis function on the dataframe x. The optionnal card argument can provide a vector of cardinalities for each attribute (i.e. column) of the dataframe. If absent, the cardinalities are computed from the dataframe, but not accurate if some possible values never show up. See papers "Scaling log-linear analysis to high-dimensional data, ICDM 2013", "Scaling log-linear analysis to datasets with thousands of variables, SDM 2015", and "A multiple test correction for streams and cascades of statistical hypothesis tests, KDD 2016" for more details.

Value

A Chordalysis object. Use ChoR.as.* functions to access the result.

Examples

## Not run:  res = ChoR.Budget(data) 
## Not run:  res = ChoR.Budget(data, budgetShare=0.0) 
## Not run:  res = ChoR.Budget(data, 0.05, card = c(3, 5, 4, 4, 3, 2, 3, 3))