multiclass_logloss: Calculate the multiclass logloss
In alexWhitworth/glmEnsemble: Builds GLM ensemble models in parallel

Calculates the multiclass logloss defined: \frac{1}{N}∑_i^N∑_j^M y_{ij} log(p_ij)

1	multiclass_logloss(p_mat, class_mat, w_vec = NULL)

`p_mat`	An N x m matrix of predicted probabilities for a multiclass problem. N == number of observations. m = number of classes. It is assumed that the row-sums all == 1
`class_mat`	An N x m 0,1 integer matrix giving the true classes of the observations. It is assumed that the row-sums all == 1.
`w_vec`	An optional vector of observation weights. Defaults to NULL ie - equal weights