✳flâneur — a map of the web's best reading

MDE - pymde 0.1.18 documentation

pymde.org · 4,133 words · saved by 1 readers

The central concept in PyMDE is the Minimum-Distortion Embedding (MDE) problem. An MDE problem is an optimization problem whose solution is an embedding. We can interpret an MDE problem as a declarative description of the properties an embedding should satisfy. MDE problems in PyMDE are represented by the pymde.MDE object. In this part of the tutorial, we explain what an MDE problem is. Then we show how to construct MDE problems using PyMDE, with custom objective functions and constraints. We also describe some of the useful methods that the pymde.MDE class provides, and how to sanity-check and compare embeddings. In this section, we introduce the concept of an MDE problem, whose solution is an embedding. At a high-level, the objective of an MDE problem is to minimally distort known relationships between some pairs of items, while possibly satisfying some constraints. We first explain this abstractly. In the next section, we show how to make MDE problems using PyMDE. An MDE problem sta

MDE ¶ The central concept in PyMDE is the Minimum-Distortion Embedding (MDE) problem. An MDE problem is an optimization problem whose solution is an embedding. We can interpret an MDE problem as a declarative description of the properties an embedding should satisfy. MDE problems in PyMDE are represented by the pymde.MDE object. In this part of the tutorial, we explain what an MDE problem is. Then we show how to construct MDE problems using PyMDE, with custom objective functions and constraints. We also describe some of the useful methods that the pymde.MDE class provides, and how to sanity-ch

Explore this link on the map →

related reading