Close the Feedback Loop

Machine learning needs data-driven feedback for active learning, and data scientists need qualitative feedback to focus new development.​

close feedback loop

What You Can Expect

Iteration
Functionality that enables iteration​
Data Set
Airline-provided decision support capabilities
Scatter Plot
Support for airline-provided protection sets​
Experimentation
A workflow & experimentation feedback loop

We Give You the Tools to Get Started

We provide the system with our Eddy platform, a user-driven workflow, a format for models and metrics, and a data fee to facilitate model development.

illustration-workflow1

Optimization

  • Airline-provided protection levels for all classes, including discrete classes
  • Support for provision of multiple sets of protection levels (akin to different profiles)
  • Store it
  • Calculate recommended AUs
  • Calculate and display LAF
  • Send to PSS
security icon

Decision Support

  • Fields airline chooses to provide based on business need
  • Intentionally flexible to match airline business process evolution
  • Display and filter in Control and Details
  • No logic applied to these
decision icon

User Guardrails & Pacing Curves

  • Airline-provided pacing curves and tolerance thresholds to allow for custom performance tracing on a per-flight-cabin basis​.
  • Plot in Details
  • Calculate Performance Bands
  • Replaces existing clusters / pacing curves workflow if enabled
guardrail icon

Forecasting

  • Forecasts provided by an airline that are designed for use with a Kambr-provided optimizer
  • Store it
  • Adjust it in the UI
  • Use as input to the optimizer
forecasting icon