iterateBMAglm_train_predict_test: Iterative Bayesian Model Averaging: training, prediction and...
In iterativeBMA: The Iterative Bayesian Model Averaging (BMA) algorithm

Description Usage Arguments Details Value Note References See Also Examples

Classification and variable selection on microarray data. This is a multivariate technique to select a small number of relevant variables (typically genes) to classify microarray samples. This function performs the training, prediction and testing steps. The data is assumed to consist of two classes, and the classes of the test data is assumed to be known. Logistic regression is used for classification.

1	iterateBMAglm.train.predict.test (train.expr.set, test.expr.set, train.class, test.class, p=100, nbest=10, maxNvar=30, maxIter=20000, thresProbne0=1)

`train.expr.set`	an `ExpressionSet` object. We assume the rows in the expression data represent variables (genes), while the columns represent samples or experiments. This training data is used to select relevant genes (variables) for classification.
`test.expr.set`	an `ExpressionSet` object. We assume the rows in the expression data represent variables (genes), while the columns represent samples or experiments. The variables selected using the training data is used to classify samples on this test data.
`train.class`	class vector for the observations (samples or experiments) in the training data. Class numbers are assumed to start from 0, and the length of this class vector should be equal to the number of rows in train.dat. Since we assume 2-class data, we expect the class vector consists of zero's and one's.
`test.class`	class vector for the observations (samples or experiments) in the test data. Class numbers are assumed to start from 0, and the length of this class vector should be equal to the number of rows in test.dat. Since we assume 2-class data, we expect the class vector consists of zero's and one's.
`p`	a number indicating the maximum number of top univariate genes used in the iterative BMA algorithm. This number is assumed to be less than the total number of genes in the training data. A larger p usually requires longer computational time as more iterations of the BMA algorithm are potentially applied. The default is 100.
`nbest`	a number specifying the number of models of each size returned to `bic.glm` in the `BMA` package. The default is 10.
`maxNvar`	a number indicating the maximum number of variables used in each iteration of `bic.glm` from the `BMA` package. The default is 30.
`maxIter`	a number indicating the maximum of iterations of `bic.glm`. The default is 20000.
`thresProbne0`	a number specifying the threshold for the posterior probability that each variable (gene) is non-zero (in percent). Variables (genes) with such posterior probability less than this threshold are dropped in the iterative application of `bic.glm`. The default is 1 percent.

This function consists of the training phase, prediction phase, and the testing phase. The training phase consists of first ordering all the variables (genes) by a univariate measure called between-groups to within-groups sums-of-squares (BSS/WSS) ratio, and then iteratively applying the bic.glm algorithm from the BMA package. The prediction phase uses the variables (genes) selected in the training phase to classify the samples in the test set. The testing phase assumes that the class labels of the samples in the test set are known, and computes the number of classification errors and the Brier Score.

A list consisting of 4 elements are returned:

`num.genes`	The number of relevant genes selected using the training data.
`num.model`	The number of models selected using the training data.
`num.err`	The number of classification errors produced when the the predicted class labels of the test samples are compared to the known class labels.
`brierScore`	The Brier Score computed using the predicted and known class labels of the test samples. The Brier Score represents a probabilistic number of errors. A small Brier Score implies high prediction accuracy.

The BMA and Biobase packages are required.

Raftery, A.E. (1995). Bayesian model selection in social research (with Discussion). Sociological Methodology 1995 (Peter V. Marsden, ed.), pp. 111-196, Cambridge, Mass.: Blackwells.

Yeung, K.Y., Bumgarner, R.E. and Raftery, A.E. (2005) Bayesian Model Averaging: Development of an improved multi-class, gene selection and classification tool for microarray data. Bioinformatics 21: 2394-2402.

iterateBMAglm.train, iterateBMAglm.train.predict

library (Biobase)
library (BMA)
library (iterativeBMA)
data(trainData)
data(trainClass)
data (testData)
data (testClass)

iterateBMAglm.train.predict.test (train.expr.set=trainData, test.expr.set=testData, trainClass, testClass, p=100)