Matrix Factorization

تحليل المصفوفات

أسلوب أساسي في الأنظمة الموصية يُحلِّل مصفوفة تفاعلات المستخدم والعنصر إلى مصفوفتَي عوامل كامنة (للمستخدمين والعناصر)، يُتنبأ بالتقييم أو التفضيل بحاصل الضرب النقطي بينهما، وذاع صيته بعد فوز فريق BellKor في تحدي نتفليكس.

A foundational recommender-system technique that decomposes the user–item interaction matrix into two latent-factor matrices for users and items, predicting a rating or preference from their dot product, popularized by the BellKor team's win in the Netflix Prize.

Also translated asMatrix Factorization، MF، تحليل المصفوفة، تفكيك المصفوفات في الأنظمة الموصية

First appears in this corpus in: GroupLens: An Open Architecture for Collaborative Filtering of Netnews (1994)

Appears in these papers

Discovering Faster Matrix Multiplication Algorithms with Reinforcement Learning2022 in the sky ✦
BPR: Bayesian Personalized Ranking from Implicit Feedback2009 in the sky ✦
BPR: Bayesian Personalized Ranking from Implicit Feedback2009 in the sky ✦
GloVe: Global Vectors for Word Representation2014 in the sky ✦
GloVe: Global Vectors for Word Representation2014 in the sky ✦
GroupLens: An Open Architecture for Collaborative Filtering of Netnews1994 in the sky ✦
GroupLens: An Open Architecture for Collaborative Filtering of Netnews1994 in the sky ✦
Amazon.com Recommendations: Item-to-Item Collaborative Filtering2003 in the sky ✦
Amazon.com Recommendations: Item-to-Item Collaborative Filtering2003 in the sky ✦
Matrix Factorization Techniques for Recommender Systems2009 in the sky ✦
Matrix Factorization Techniques for Recommender Systems2009 in the sky ✦
Neural Collaborative Filtering2017 in the sky ✦
Neural Collaborative Filtering2017 in the sky ✦
Deep Neural Networks for YouTube Recommendations2016 in the sky ✦