Code for predicting missing elements with an existing CMF model.
The predictions are made for all of the elements specified in the list of
input matrices X. The function also returns the root mean square error
(RMSE) between the predicted outputs and the values provided in X.
Usage
predictCMF(X, model)
Value
A list of
out
A list of matrices corresponding to predictions for each
matrix in X.
error
A vector containing the root-mean-square error for each
matrix separately.
Arguments
X
A list of sparse matrices specifying the indices for which to
make the predictions.
These matrices must correspond to the structure used for X
when learning the model with CMF.
model
A list of model parameter values provided by CMF.
Author
Arto Klami and Lauri Väre
Details
Note that X needs to be provided as a set of triplets instead of as
a regular matrix. See matrix_to_triplets().