awesome-repositories.com
Blog
awesome-repositories.com

Découvrez les meilleurs dépôts open-source grâce à notre recherche par IA.

ExplorerRecherches sélectionnéesAlternatives open sourceLogiciels auto-hébergésBlogPlan du site
ProjetÀ proposNotre méthodologiePresseServeur MCP
Mentions légalesConfidentialitéConditions d'utilisation
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
·
lyst avatar

lyst/lightfm

0
View on GitHub↗
5,095 stars·724 forks·Python·Apache-2.0·13 vues

Lightfm

LightFM est une bibliothèque de recommandation Python et un framework de machine learning conçu pour prédire les préférences des utilisateurs. Il implémente un moteur de recommandation hybride qui combine le filtrage collaboratif avec le filtrage par contenu en intégrant les données d'interaction utilisateur-élément avec des métadonnées descriptives.

Le système utilise la factorisation de matrice hybride pour apprendre des représentations latentes des utilisateurs et des éléments. Il est spécifiquement conçu pour gérer les feedbacks implicites, utilisant des fonctions de perte spécialisées telles que Weighted Approximate Rank Pairwise (WARP) et Bayesian Personalized Ranking (BPR) pour optimiser les préférences d'éléments pour les jeux de données manquant de notes négatives.

La bibliothèque fournit des outils pour entraîner des modèles via la descente de gradient stochastique, calculer des prédictions de préférence d'éléments et évaluer la précision du modèle. Elle prend en charge le classement personnalisé des éléments et la prédiction du comportement des utilisateurs en synthétisant des matrices d'interaction avec des plongements de caractéristiques (embeddings).

Features

  • Recommender Systems - Provides a hybrid recommendation system combining collaborative filtering with content-based metadata.
  • Interaction Matrix Factorizers - Uses latent-factor matrix factorization to decompose user-item interactions into shared embedding spaces.
  • Hybrid Matrix Factorization - Learns latent representations of users and items using hybrid matrix factorization.
  • Recommendation Engines - Implements a hybrid recommendation engine that integrates interaction data with user and item metadata.
  • Implicit Feedback Models - Optimizes item preferences using WARP and BPR loss functions for datasets without negative ratings.
  • Preference Prediction - Calculates preference scores for users and items based on learned latent representations.
  • Implicit Feedback Modeling - Designed to handle implicit feedback using specialized loss functions for datasets lacking negative ratings.
  • Recommendation List Generators - Generates ranked lists of item suggestions by synthesizing interaction data and metadata.
  • Hybrid Feature Embeddings - Combines collaborative filtering with item and user metadata by summing their latent vectors.
  • Personalized Item Ranking - Calculates preference scores for specific user-item pairs to generate personalized rankings.
  • Hybrid Recommendation Training - Learns preferences by processing interaction matrices and metadata via stochastic gradient descent.
  • Recommendation Model Training - Optimizes model parameters using interaction data and specific loss functions to learn user behavior.
  • Model Performance Evaluators - Includes tools for evaluating model precision and accuracy by comparing suggestions against ground truth.
  • Implicit Feedback Optimization - Trains models on positive-only interaction datasets using specialized losses like WARP or BPR.
  • Top-K Accuracy Evaluators - Measures recommendation accuracy by evaluating the proportion of relevant items in the top-k results.
  • Recommendation Libraries - Provides a comprehensive Python library for training and evaluating recommendation models.
  • User Behavior - Learns patterns from interaction matrices and metadata to predict future user behavior.
  • Stochastic Gradient Descent - Utilizes stochastic gradient descent to iteratively update model weights based on interaction samples.
  • Personalized Ranking Optimizers - Implements Bayesian Personalized Ranking to maximize the probability of user preference for interacted items.
  • Recommender Frameworks - Python implementation of collaborative and content-based learning-to-rank algorithms.
  • Recommender Systems - Hybrid recommendation algorithms for implicit and explicit feedback.

Historique des stars

Graphique de l'historique des stars pour lyst/lightfmGraphique de l'historique des stars pour lyst/lightfm

Recherche par IA

Explorez plus de dépôts awesome

Décrivez vos besoins en langage naturel — l'IA classe des milliers de projets open source sélectionnés par pertinence.

Start searching with AI

Alternatives open source à Lightfm

Projets open source similaires, classés selon le nombre de fonctionnalités partagées avec Lightfm.
  • benfred/implicitAvatar de benfred

    benfred/implicit

    3,797Voir sur GitHub↗

    Implicit is a Python recommendation engine and matrix factorization library designed for collaborative filtering. It implements predictive models that analyze implicit feedback to estimate user preferences and generate personalized item recommendations without requiring explicit ratings. The library utilizes native-code execution and multi-core parallelized processing to decompose large interaction matrices into latent factors. It incorporates approximate nearest neighbor indexing to accelerate high-dimensional similarity lookups and reduce recommendation latency. The framework covers prefer

    Pythoncollaborative-filteringmachine-learningmatrix-factorization
    Voir sur GitHub↗3,797
  • nicolashug/surpriseAvatar de NicolasHug

    NicolasHug/Surprise

    6,793Voir sur GitHub↗

    Surprise is a Python library for building and analyzing recommendation systems. It provides a comprehensive toolkit for implementing collaborative filtering to predict user preferences and generate item suggestions based on historical rating patterns. The library includes dedicated tools for hyperparameter optimization and model evaluation. It allows for searching through parameter sets to find the most effective configurations and utilizes a suite of metrics to measure prediction accuracy. The framework covers the full development workflow, including data loading from various sources, the c

    Pythonfactorizationmachine-learningmatrix
    Voir sur GitHub↗6,793
  • gorse-io/gorseAvatar de gorse-io

    gorse-io/gorse

    9,717Voir sur GitHub↗

    Gorse is a personalized recommendation engine server and machine learning pipeline designed to suggest items to users based on their behavior and preferences. It operates as a distributed system that separates training, candidate generation, and serving nodes to support high-throughput workloads. The system utilizes a multi-stage recommendation pipeline to refine results through retrieval, scoring, and reranking. It generates personalized suggestions using collaborative filtering, matrix factorization, and item-to-item similarity models, while also providing non-personalized and fallback reco

    Gocollaborative-filteringgoknn
    Voir sur GitHub↗9,717
  • recommenders-team/recommendersAvatar de recommenders-team

    recommenders-team/recommenders

    21,769Voir sur GitHub↗

    This project is a recommendation system framework designed for building, evaluating, and operationalizing personalized item suggestion engines. It provides a comprehensive toolkit for implementing collaborative filtering and content-based algorithms, supported by an end-to-end machine learning pipeline for preparing datasets and deploying predictive models. The framework distinguishes itself through the integration of knowledge graphs to provide richer context for recommendations and the use of industry-specific patterns to accelerate system deployment. It also includes a specialized model ev

    Pythonaiartificial-intelligencedata-science
    Voir sur GitHub↗21,769
Voir les 30 alternatives à Lightfm→

Questions fréquentes

Que fait lyst/lightfm ?

LightFM est une bibliothèque de recommandation Python et un framework de machine learning conçu pour prédire les préférences des utilisateurs. Il implémente un moteur de recommandation hybride qui combine le filtrage collaboratif avec le filtrage par contenu en intégrant les données d'interaction utilisateur-élément avec des métadonnées descriptives.

Quelles sont les fonctionnalités principales de lyst/lightfm ?

Les fonctionnalités principales de lyst/lightfm sont : Recommender Systems, Interaction Matrix Factorizers, Hybrid Matrix Factorization, Recommendation Engines, Implicit Feedback Models, Preference Prediction, Implicit Feedback Modeling, Recommendation List Generators.

Quelles sont les alternatives open-source à lyst/lightfm ?

Les alternatives open-source à lyst/lightfm incluent : benfred/implicit — Implicit is a Python recommendation engine and matrix factorization library designed for collaborative filtering. It… nicolashug/surprise — Surprise is a Python library for building and analyzing recommendation systems. It provides a comprehensive toolkit… gorse-io/gorse — Gorse is a personalized recommendation engine server and machine learning pipeline designed to suggest items to users… recommenders-team/recommenders — This project is a recommendation system framework designed for building, evaluating, and operationalizing personalized… rucaibox/recbole — RecBole is a PyTorch-based recommendation framework designed for building, training, and evaluating a wide variety of… d2l-ai/d2l-en — This project is an educational platform and research toolkit designed to teach deep learning through a combination of…