Skip to main content


Project description

# ALS-WR : Alternating-Least-Squares with Weighted-λ-Regularization

This is the code to perform ALS-WR as presented by [Zhou et al]

The code is mainly inspired from [Mendeley Ltd]

- *R* ∈ ℝ<sup>*n* × *p*</sup> = (*r*<sub>*i**j*</sub>) denote the sparse user-item matrix.
- 𝒟 is the set of available ratings.
- Assume a low rank constraint *K* on the matrix : *R* = *U**V*<sup>*T*</sup>, *U*, *V* ∈ ℝ<sup>*n* × *K*</sup>, ℝ<sup>*p* × *K*</sup>
- *U*<sub>*i*</sub> and *V*<sub>*j*</sub> denotes respectively the latent features of user *i* and item *j*.
- *n*<sub>*u*<sub>*i*</sub></sub> is the number of ratings provided by user *i*.
- *n*<sub>*v*<sub>*j*</sub></sub> is the number of ratings available for item *j*.
- *J*<sub>*i*</sub> and *L*<sub>*j*</sub> is respectively the set of rated item by user *i* and available ratings for item *j*.

Objective function :
*f*(*U*, *V*)=∑<sub>*i*, *j* ∈ 𝒟</sub>(*r*<sub>*i**j*</sub> − *U*<sub>*i*</sub>*V*<sub>*j*</sub><sup>*T*</sup>)<sup>2</sup> + *λ*(∑<sub>*i*</sub>*n*<sub>*u*<sub>*i*</sub></sub>||*U*<sub>*i*</sub>||<sup>2</sup>+∑<sub>*i*</sub>*n*<sub>*v*<sub>*j*</sub></sub>||*V*<sub>*j*</sub>||<sup>2</sup>)


- Input : *R*, *K* (possibly an initialization of *U* or *V* with SVD)
- Until convergence do
- For each user (in parallel if wanted)
- *U*<sub>*i*</sub> = (*V*<sub>*I*<sub>*i*</sub></sub>*V*<sub>*I*<sub>*i*</sub></sub> + *λ**I*<sub>*K*</sub>)<sup>−1</sup>*V*<sub>*I*<sub>*i*</sub></sub>*R*<sub>*i*, *I*<sub>*i*</sub></sub>
- For each item (in parallel if wanted)
- *V*<sub>*j*</sub> = (*U*<sub>*L*<sub>*j*</sub></sub>*U*<sub>*L*<sub>*j*</sub></sub> + *λ**I*<sub>*K*</sub>)<sup>−1</sup>*U*<sub>*L*<sub>*j*</sub></sub>*R*<sub>*L*<sub>*j*</sub>, *j*</sub>

[Zhou et al]:
[Mendeley Ltd]:

Project details

Release history Release notifications

This version
History Node


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
pyalswr-1.0.0-py2.py3-none-any.whl (3.0 kB) Copy SHA256 hash SHA256 Wheel py2.py3 Mar 26, 2016
pyalswr-1.0.0.tar.gz (2.5 kB) Copy SHA256 hash SHA256 Source None Mar 26, 2016

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging CloudAMQP CloudAMQP RabbitMQ AWS AWS Cloud computing Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page