MEfunctions {integrativeME} | R Documentation |
Different internal gating functions (or models) are proposed within mixture of experts to integrate gene expression data and clinical data in a binary classification framework.
MEindep(jcross, train, test, n, nq, nv, ng, indclass, data.cat, data.cont, prop.kmeans, means.kmeans, var.kmeans) MElogreg(jcross, train, test, n, nq, nv, ng, indclass, data.cat, data.cont, prop.kmeans) MEloc(jcross, train, test, n, nq, nv, ng, indclass, data.cat, data.cont, prop.kmeans, means.kmeans, var.kmeans, loc.ind)
jcross |
which cross validation sample |
train |
training samples |
test |
test samples |
n |
number of observations or samples |
nq |
number of clinical variables |
nv |
number of genes, the genes should be selected beforehand, see integrativeME . |
ng |
number of experts, should be set to 2 for a binary classification problem. |
indclass |
number of samples of class 0. |
data.cat |
clinical data (categorical). |
data.cont |
gene expression data. |
prop.kmeans |
proportions, initialized with k-means, see also kmeans.init . |
means.kmeans |
means, initialized with k-means, see also kmeans.init . |
var.kmeans |
variance-covariance matrix, initialized with k-means, see also kmeans.init . |
loc.ind |
index of the location variable in the case of the MEloc model. |
Given a training set and a test set, the parameters in integrativeME are learnt via the EM algorithm and then tested. All three geting functions are included in the main program integrativeME
prop |
estimated proportions. |
w |
weighted variable vector in the expert networks function. |
loglik |
loglikelihood of the model after several iterations. |
mat.gum |
main output that is used in integrativeME to predict the class label of each tested observation. |
Kim-Anh Le Cao
Le Cao et al. (2009), submitted.
Ng, S.K. and McLachlan, G.J. (2008). Expert Networks with Mixed Continuous and Categorical Feature Variables: a Location Modeling Approach. Machine Learning Research Progress, ed. Hanna Peters and Mia Vogel, 1–14
Hunt, L. and Jorgensen, M. (1999). Mixture model clustering using the MULTIMIX program. Australian & New Zealand Journal of Statistics, 41, 2, 154–171.