Conditional independence tests were performed using a shrinkage mutual information test [33] with α = 0.05.
The mutual information test calculates the probability of the observed mutual information to be higher than the random mutual information.
Such a test, unlike the more common asymptotic χ mutual information test, is valid and has been shown to work reliably even on small samples.
This is equivalent to assessing the mutual information and testing for significant departures from zero.
Relationships between eRSN characteristics and outcome will be sought using approaches including pair-wise relationships such as mutual information measures, with testing using surrogate signals as well as different statistical testings at individual and group levels.
A co-transcriptional network was constructed from the microarray data using the ARACNE method, which uses mutual information (MI) to test statistical dependency between expression profiles [ 11].
Additionally, by comparing experimental conditions with the same number of feedback events (sounds), but varying in the amount of information feedback conveys about outcome (the mutual information), we can test if information content has an effect on performance.
Specifically, we used mutual information as the test statistic to calculate how unlikely it is to get a mutual information at least as large as the observed one under the null hypothesis that the two sites are independent.
Using similar methods, we defined common information theory parameters including entropy and mutual information for common lab tests done concurrently or sequentially, without reference to a final diagnosis.
Theoretically, any wrapper or filter feature selection method, such as t-test, mutual information measurement, etc, can be used.
We used constructs from information theory, namely entropy, conditional entropy, and mutual information, to analyze lab test results from a large ICU database.
