GetMaximumDValueFromTwoByTwoContingencyTable: get maximum D value from two-by-two contingency table

Description Usage Arguments Value References

View source: R/GetMaximumDValueFromTwoByTwoContingencyTable.r

Description

Computes the maximum D value (or weight) given approximate values of n++, n+-, n-+ and n–. These values are approximate and in general are non-integer values; they are found by using an approximation detailed in the paper Assessing statistical significance in causal graphs on page 6 i.e. n_ab is approximately equal to q_a*n_b/t where a and b are either +, - or 0. The value is an approximation since the direction in which the number should be rounded is not clear and hence this function runs through all possible combinations of rounding before concluding the maximum D-value.

Usage

1
2
3
GetMaximumDValueFromTwoByTwoContingencyTable(twoByTwoContingencyTable,
  predictionListStats, experimentalDataStats,
  logOfFactorialOfPredictionListStats, returnlog = FALSE)

Arguments

twoByTwoContingencyTable

approximate values of n++, n+-, n-+ and n–, these values arecalculated to optimise the D-value

predictionListStats

a vector containing the values q+, q- and q0 the number of positive/negative/non-significant (or contradictory) predictions)

experimentalDataStats

a vector containing the values n+, n- and n0 (the number of positive/negative/non-significant (or contradictory) transcripts in the results)

logOfFactorialOfPredictionListStats

a vector containing the log of the factorial value for each entry in predictionListStats

returnlog

whether or not he value should be returned as a log (TRUE) or not (FALSE)

Value

the maximal D-value

References

L Chindelevitch et al. Assessing statistical significance in causal graphs. BMC Bioinformatics, 13(35), 2012.


CausalR documentation built on Nov. 8, 2020, 5:25 p.m.