Skip to content

Introduction

The KX ML Analytics Library acts as a centralized location for all machine learning analytics produced by KX. This is an expansion of the KX ML Toolkit and is intended to provide users with a package which can be integrated into the kdb Insights Enterprise and Microservices.

The analytics described here cover the following areas:

  1. Data preprocessing
  2. Machine learning models
  3. Model/function optimisation
  4. Model evaluation

Functionality breakdown

ML Toolkit

Functionality present in the ML Toolkit has been included within the ML Analytics package described here. The functionality includes, but is not limited to:

  • Preprocessing functions
  • Clustering models
  • Time-series models

Documentation can be found at code.kx.com and should be used in conjunction with the additional documentation found within this package.

ML Analytics

In addition to the ML Toolkit functionality, ML Analytics contains the following:

  • Online learning/out-of-core algorithms
  • Variadic function signatures

Sections

Documentation follows the structure below:

Note

The following section has been omitted in the move from the ML Toolkit to ML Analytics:

This section has been omitted due to potential conflation between the concept of a pipeline as it pertains to the Stream Processor and graph use within the ML Toolkit.