Infer.NET user guide

Learners

A “learner” is a complete solution for a machine learning application such as classification or recommendation. Learners can be invoked directly from the command-line or a .NET program without having to learn the Infer.NET API. Each learner includes training. prediction, and evaluation capabilities. The source code for learners is included in the Infer.NET release to serve as comprehensive examples of how to build robust machine learning functionality using Infer.NET.

• Bayes Point Machine classifiers

  • Introduction
  • Learner API
    • Mappings
      • Standard data format mapping
      • Native data format mapping
      • Evaluation data format mapping
    • Creation & serialization
    • Settings & events
    • Training
    • Prediction
    • Evaluation
  • Model
  • Command-line runners
    • Training
    • Incremental training
    • Prediction
    • Cross-validation
    • Convergence diagnosis
    • Sampling weights
    • Evaluation

• Matchbox recommender

  • Introduction
  • Learner API
    • Data mappings
      • Standard data format mapping
      • Native data format mapping
      • Data splitting mapping
      • Negative data generation mapping
    • Setting up a recommender
    • Training
    • Prediction
    • Evaluation
      • Rating prediction
      • Item recommendation
      • Related users
      • Related items
    • Serialization
  • Model
  • Command-line runners
    • Negative data generator
    • Data splitter
    • Trainer
    • Predictors
    • Evaluators