bayesian.classifier

function (data, method, training, explanatory, whitelist, blacklist,
    expand, debug = FALSE)
{
    check.logical(debug)
    check.learning.algorithm(method, class = "classifier")
    check.nodes(training, max.nodes = 1)
    check.data(data, allowed.types = discrete.data.types)
    if (missing(data)) {
        check.nodes(explanatory)
    }
    else {
        vars = names(data)
        check.nodes(training, graph = vars, max.nodes = 1)
        if (missing(explanatory))
            explanatory = vars[vars != training]
        else check.nodes(explanatory, graph = nodes)
    }
    if (training %in% explanatory)
        stop("node ", training, " is included in the model both as a training ",
            "and an explanatory variable.")
    nodes = c(training, explanatory)
    if (method != "naive.bayes") {
        whitelist = build.whitelist(whitelist, nodes, data, criterion = "mi")
        blacklist = build.blacklist(blacklist, whitelist, nodes)
        if (method == "tree.bayes") {
            if ((training %in% whitelist) || (training %in% blacklist))
                stop("blacklisting arcs to and from the training node is not allowed.")
        }
    }
    extra.args = check.classifier.args(method = method, data = data,
        extra.args = expand, training = training, explanatory = explanatory)
    if (method == "naive.bayes") {
        ntests = 0
        test = "none"
        res = naive.bayes.backend(data = data, training = training,
            explanatory = explanatory)
    }
    else if (method == "tree.bayes") {
        ntests = length(explanatory) * (length(explanatory) -
            1)/2
        test = as.character(mi.estimator.tests[extra.args$estimator])
        res = tan.backend(data = data, training = training, explanatory = explanatory,
            whitelist = whitelist, blacklist = blacklist, mi = extra.args$estimator,
            root = extra.args$root, debug = debug)
    }
    res$learning$algo = method
    res$learning = list(whitelist = whitelist, blacklist = blacklist,
        test = test, ntests = ntests, algo = method, args = extra.args,
        optimized = NULL)
    res$learning$args$training = training
    invisible(res)
}