QMI: Compute different mutual information (MI) theoretic measures...
In sealx017/MIAMI: Mutual Information-based Analysis of Multiplex Imaging data

Compute different mutual information (MI) theoretic measures such as EQMI*, EQMI and CSQMI in a general dataset

1	QMI(dat, bandwidth = "HPI", measure = "EQMI_star", var_names = F)

`dat`	is the data matrix whose p columns correspond to p different random variables (markers in a multiplex imaging dataset) and whose rows are samples (usually, different cells from different subjects in a multiplex imaging dataset).
`bandwidth`	is the bandwidth selection procedure user wants to use. The default option "HPI" corresponds to multivariate plug-in bandwidth estimates. For faster computation, use "Ind" which would independently select bandwidth for every r.v. by using Silverman's rule.
`measure`	corresponds to the MI measure one wants to compute. For example, measure = "EQMI_star" will output the measure EQMI_star. Other options are, "EQMI, "CSQMI" and "ALL", where the last one outputs all the three measures.
`var_names`	can be `TRUE` (T) or `FALSE` (F). If it is True the output list will have name of the measures.

It returns a list whose first element is the vector of estimated MI between the following combination of the markers, (1, 2), (1, 2, 3), (1, 2, 3, 4), ..., and (1, 2, ..., p), and the second element is the vector of bandiwidth parameters used. If measure = "All", the first element is a list of three elements corresponding to EQMI_star, EQMI and CSQMI.

sealx017/MIAMI documentation built on Feb. 11, 2022, 12:20 a.m.