Skip to main content

What is Skeye Predict™?

Skeye Predict is one workspace for building, validating, and shipping interpretable prediction models — models made of rules you can read, question, and explain, rather than a black box.

Sign in to Skeye Predict

Three modules​

Predictor — rule-based models you can explain​

Train interpretable, rule-based prediction models. Tune precision and coverage, run k-fold cross-validation, and inspect the rules before you ship.

→ Running a prediction

EEDA — see where your model wins and fails​

The Easy Exploratory Data Analyser. You're the business expert; EEDA helps you test your hypothesis rules against the data and see, live, where they hold up.

→ EEDA

Knowledge Base — one home for your validated rules​

Promote validated automatic and expert rules into a shared library your whole team can reuse.

→ Knowledge Base

The typical workflow​

  1. Create a project from a data file or a Skeye Ingest folder.
  2. Run a prediction — set the parameters, launch the run, explore the model, and export it as .pkl or .pmml for your downstream systems.
  3. Explore with EEDA to see where the model performs, and turn good patterns into rules.
  4. Promote rules to the Knowledge Base so the team can reuse them.
  5. Test your model on new inputs, and review past runs in Runs History.