A measure of association is computed using a ratio of
maximum likelihood scores for two distributions, joint and marginal.
Both likelihoods are maximized independently with respect to their kernel width.
Usage
ma(d,partition,ht,hp,hs)
Arguments
Value
Returns a list of values ...Aa score (including hyperbolic correction) estimating association for the datarawAthe association score before hyperbolic correctionjointKWthe optimal kernel width for the joint distributionaltLLthe optimal weighted log likelihood for the alternate distributionnullLLthe optimal log likelihood for the marginal distributionmarginalKWthe optimal kernel width for the marginal distributionweightthe optimal weight used for the mixtureLRstatthe LR statistic, required for computing p values.nRowsn, the number of complete samples in the data setmColsm, the number of variables in the data setpartitionuser supplied partition for the variables in the data set
Details
An estimate of association (possibly nonlinear) is computed
using a ratio of
maximum likelihoods for the marginal distribution and
maximum weighted likelihoods for the joint distribution.
Before the computation is carried out the data is ranked using the
rwt function from the matie package.
This estimate is usually conservative (ie low) and a hyperbolic
correction is applied by adding an offset, os,
to the joint likelihood given by:
latex{
$\code{os} = \left( 1 - \frac{1}{1 + \code{A} \times \code{ht}} \right)
\frac{\code{n}^{\code{hp}}}{ \code{hs} }$
}
html{
os = ( 1 - 1 / (1 + A * ht) ) * ( n ^ (hp) / hs )
}
before the ratio is re-computed.