EC2 Backend

Features

• Added the ability to power your local Prediction Lab instance with your own EC2 backend.
• Improved the memory pressure across the application.
• Optimized the automated model build creation, improving performance by up to 75% on large datasets.

Bring Your Own Cloud

Features

• Added support for Bring Your Own Cloud.

Single Tenant Deployment

Features

• Added support for single-tenant deployment.

Activity Panel

Features

• Added a new activity panel for datasets and models.

Coefficient Table

Features

• Added a new coefficient table view.

Model Effects

Features

• Added the ability to visualize and break apart individual model effects

Command Palette

Features

• Added a keyboard first command palette for application navigation

Model Reviews

Features

• Added the ability to create reviews of your datasets
• Added the ability to create reviews of your models

Model Import

Features

• Added support for importing models from legacy software

Two-Way Actual vs Predicted

Features

• Added a new two-way actual vs predicted view

Model Partitions

Features

• We now support fitting models with partitions

Realtime Support

Features

• You can now submit feature requests directly within the app
• You can now submit bug reports directly within the app, giving our support team all of the necessary information to reproduce the issue and fix it quickly
• You can now chat live with our support team or immediately jump into a video call to discuss anything predictive modeling related

Roles and Permissions

Features

• Added the ability to define roles and assign permissions at the granular level
• Added the ability to assign roles to users and agents

Multi Tenant Deployment

Features

• Added support for multi-tenant deployment.

SOC II Type I Compliance

Features

• Announcing SOC II Type I compliance.

Error Structures and Link Functions

Features

• We now support Gaussian, Binomial, Poisson, Gamma, Inverse Gaussian, and Tweedie error structures
• We now support Identity, Log, Inverse, Inverse-square, Square-root, Logit, Probit, Log-log, and Complementary log-log link functions

Automated Feature Creation

Features

• Automate your feature creation process by fitting an initial Gradient Boosting Machine

Lift Chart

Features

• Added a new lift chart view

Version Control

Features

• Added the ability to create branches of your dataset
• Added the ability to create branches of your models

Correlation View

Features

• Added a new correlation view

Dataset Upload

Features

• Support for .parquet and .csv files

Initial Launch

Prediction Lab is a predictive modeling platform that focuses on speed to insight.