Find a model

7 models· Recommendation Systems, Time-Series Forecasting

—

by Meta · 0 reviews

An open-source deep learning architecture for recommendation systems that combines dense and sparse features through embedding tables, interaction layers, and MLP components, used for CTR prediction and personalization.

Accuracy—

Speed—

Cost

—

Prophet

by Meta (Facebook) · 0 reviews

An open-source time series forecasting library that uses an additive model with non-linear trends, seasonal components, and holiday effects, available in Python and R.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare

—

Temporal Fusion Transformer

by Google Research · 0 reviews

A transformer-based architecture for multi-horizon time series forecasting that provides interpretable predictions through attention mechanisms.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare

—

N-BEATS

by ServiceNow / Element AI · 0 reviews

A neural basis expansion analysis model for time series forecasting using deep fully-connected layers and residual connections.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare

—

StatsForecast

by Nixtla · 0 reviews

A high-performance time series forecasting library featuring AutoARIMA, ETS, and other statistical models optimized for speed.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare

—

libFM

by Steffen Rendle · 0 reviews

A supervised learning implementation of factorization machines for classification, regression, and ranking that captures feature interactions, widely used in recommendation systems.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare

—

DeepFM

by Huawei Noah's Ark Lab · 0 reviews

An end-to-end wide-and-deep learning framework combining factorization machines and neural networks for click-through-rate prediction.

Accuracy—

Speed—

Cost—

Ease of use—

Reliability—

Compare