Data associated with the Experiments

See also tech report (currently password protected).

alarm.cmdline - The command line used to run the experiments
runtime.log - Time required to run everything
alarm.samples - Line k contains 10000 samples from the query response distribution for the k^th query
alarm.model - Line k contains the mean query response, the estimated variance of the query response (using MeanVar), and the parameters of the Beta model for the k^th query
alarm.net - The gold standard network, only the mean parameters are used to sample data.
alarm_sampled.net - The network learnt using 500 samples from the true distribution
queries/hc91/alarm.q* - The 100 alarm queries
Log likelihoods - Log likelihood for each of the 100 alarm queries
Actual queries for HailFinder and Insurance

Network format The network format is a simplified* version of the .net format used by by LibB (Friedman and Elidan). Alarm comes from the Bayesian Network Repository (See also Friedman et al. Challenge: Where is the impact of Bayesian networks in learning? IJCAI-97). Parameters are scaled by 100 to yield Dirichlet row distributions. This doesn't affect the samples, which are generated from a network with mean parameters. A network starts by defining variables, for example,

var HISTORY FALSE TRUE defines variable HISTORY taking on two values. Then the conditional probability tables are defined, for example,

parents HISTORY LVFAILURE
FALSE  20.000000 5.000000
TRUE  8.000000 471.000000

The first variable is the child, and the following variables are parents. In this case P(HISTORY|LVFAILURE) where the first row is the distribution P(HISTORY|LVFAILURE=FALSE) and the second row is P(HISTORY|LVFAILURE=TRUE). The first value of each row corresponds to HISTORY=FALSE; the second value of each row corresponds to HISTORY=TRUE.

* The difference between LibB's format and ours is that we ignore lisp quotes and brackets, using linebreaks and spacing instead.

Query format

The following is an example of a query

q STROKEVOLUME HIGH
o HYPOVOLEMIA TRUE
o ERRCAUTER TRUE

Each line that starts with 'q' is a query variable, each line that starts with 'o' is an evidence variable. This query is

	  P(STROKEVOLUME = HIGH | HYPOVOLEMIA = TRUE, ERRCAUTER = TRUE)