The function maxsel.test computes the probability that the maximally selected chi-square statistic is <= than the value observed from the data, under the null-hypothesis of no association between X and Y, given the proportions of observations with X=1,...,X=K. The candidate binary splits over which the chi-square statistic is maximized is specified by type. If x denotes the output of the function maxsel.asymp.test, 1-x may be seen as the p-value of an independence test.
(Package: SNPmaxsel) :
Asymptotical distribution of the maximally selected chi-square statistic
This function computes the asymptotical distribution of the maximally selected chi-square statistic, where maximal selection is performed over the test statistics defined by type or user-defined through groups.
transf.inter
(Package: SNPmaxsel) :
Transforms a pair of SNPs into a single variable with nine categories
The function transf.inter creates a new variable with 9 categories out of a pair of variables with 3 categories (e.g. SNPs). The new variable is generated as described in Boulesteix et al (2007): =1 if x1=1 and x2=1, =2 if x1=2 and x2=1, =3 if x1=3 and x2=1, =4 if x1=1 and x2=2, =5 if x1=2 and x2=2, =6 if x1=3 and x2=2, =7 if x1=1 and x2=3, =8 if x1=2 and x2=3, =9 if x1=3 and x2=3.