Machine learning suite for time series

 
 

The Datapred modeling stack

Datapred is an industry-strength machine learning suite for time series. Datapred secures the 3 stages of time series modeling — preprocessing, modeling, and processing — providing unparalleled speed, flexibility and performance for time-dependent challenges.

explore-1
Explore

Experiment in our foolproof environment.

install-1
Install

Boost your operations with our modeling engine.

embed-1
Embed

Put a sequential tiger in your products.

shutterstock_530971462

Explore

Explore is the SaaS version of Datapred for data scientists. It provides a comprehensive Python framework for rapidly building prototypes that scale, while avoiding the traditional pitfalls of time series modeling. Explore’s built-in predictive models will deliver robust results in seconds, and its advanced modeling capabilities will help you achieve maximum performance in days.

Learn more

Install

Install is the algorithmic engine that, connected to your existing systems, will leverage the time series your business is already generating. Install packs the full power of the Datapred suite, and talks to most machine learning platforms, ERPs and data visualization tools. You can install it in your corporate datacenter, in a cloud you control or a public cloud. We will help you calibrate it for maximum performance on your time-dependent analysis, optimization or prediction challenge.

Learn more
shutterstock_667760623
shutterstock_668203006

Embed

Embed is the OEM version of Datapred. Add it to your products and services to provide Analytics-as-a-Service to your clients – whenever time series are involved. You can install Datapred Embed on a server and (for example) feed its outputs to a service portal, or directly on the machines that you sell… Provided the potential is there, we will adapt production architecture and pricing model to your situation. The openness of the Datapred suite ensures your control over the end product.

Learn more

Features

  • Data sequentialization
  • Stationarization and filtering
  • Built-in models
  • Compatibility with ML libraries
  • Aggregation and stacking
  • Custom cost functions
  • Parallelization and distribution
  • Connectors
  • Graph structure
  • Performance monitoring
  • Checkpoints and backups
  • Continual improvement