topN.new: Rank top-N highest-predicted items for a new user

In poismf: Factorization of Sparse Counts Matrices Through Poisson Likelihood

Rank top-N highest-predicted items for a new user

Description

Rank top-N highest-predicted items for a new user

Usage

topN.new(
  model,
  X,
  n = 10,
  include = NULL,
  exclude = NULL,
  output_score = FALSE,
  l2_reg = model$l2_reg,
  l1_reg = model$l1_reg,
  weight_mult = model$weight_mult,
  maxupd = max(1000L, model$maxupd),
  nthreads = parallel::detectCores()
)

Arguments

model	A Poisson factorization model as returned by 'poismf'.
X	Data with the non-zero item indices and counts for this new user. Can be passed as a sparse vector from package 'Matrix' ('Matrix::dsparseVector', which can be created from indices and values through 'Matrix::sparseVector'), or as a 'data.frame', in which case will take the first column as the item/column indices (numeration starting at 1) and the second column as the counts. If 'X' passed to 'poismf' was a 'data.frame', 'X' here must also be a 'data.frame'.
n	Number of top-N highest-predicted results to output.
include	List of items which will be ranked. If passing this, will only make a ranking among these items. If 'X' passed to 'poismf' was a 'data.frame', must match with the entries in its second column, otherwise should match with the columns of 'X' (numeration starting at 1). Can only pass one of 'include' or 'exclude.' Must not contain duplicated entries.
exclude	List of items to exclude from the ranking. If passing this, will rank all the items except for these. If 'X' passed to 'poismf' was a 'data.frame', must match with the entries in its second column, otherwise should match with the columns of 'X' (numeration starting at 1). Can only pass one of 'include' or 'exclude'. Must not contain duplicated entries.
output_score	Whether to output the scores in addition to the IDs. If passing 'FALSE', will return a single array with the item IDs, otherwise will return a list with the item IDs and the scores.
l2_reg	Strength of L2 regularization to use for optimizing the new factors.
l1_reg	Strength of the L1 regularization. Not recommended.
weight_mult	Weight multiplier for the positive entries over the missing entries.
maxupd	Maximum number of TNCG updates to perform. You might want to increase this value depending on the use-case.
nthreads	Number of parallel threads to use.

Details

This function calculates the latent factors in the same way as 'factors.single' - see the documentation of factors.single for details.

Just like topN, it does not exploit any potential sparsity in the fitted matrices and vectors, so it might be a lot faster to produce the recommendations externally (see the documentation for topN for details).

The factors are initialized to the mean of each column in the fitted model.

Value

• If passing 'output_score=FALSE' (the default), will return a vector of size 'n' with the top-N highest predicted items for this user.If the 'X' data passed to 'poismf' was a 'data.frame', will contain the item IDs from its second column, otherwise will be integers matching to the columns of 'X' (starting at 1). If 'X' was passed as 'data.frame', the entries in this vector might be coerced to character regardless of their original type.

• If passing 'output_score=TRUE', will return a list, with the first entry being the vector described above under name 'ix', and the second entry being the associated scores, as a numeric vector of size 'n'.

See Also

factors.single topN

poismf documentation built on March 31, 2023, 7:38 p.m.