get.all.predictions: Predict whole input matrix

Description Usage Arguments Value

Description

Outputs the predictions for the whole input matrix to which the model was fit. Note that this will be a dense matrix, and in typical recommender systems scenarios will likely not fit in memory.

Usage

1

Arguments

model

A Poisson factorization model as output by function 'poismf'.

Value

A matrix [dimA x dimB] with the full predictions for all rows and columns. If the 'X' data passed to 'poismf' was a 'data.frame', the resulting output will have row and column names added to it. Be aware that, if 'X' was a 'data.frame' with integers as IDs, selecting columns or rows of the named matrix will output ordinal positions - e.g. 'out[2,]' (second row) vs. 'out["2",]' (row for ID=2 in 'X'). The mappings, if produced, can be obtained through function get.model.mappings.


poismf documentation built on Jan. 13, 2021, 6:46 a.m.