1. Home
  2. Docs
  3. LunchBox Documentation
  4. Lunchbox ML
  5. Binary Classifier Trainer and Tester

Binary Classifier Trainer and Tester

Introduction

Binary Classification is a type of machine learning wherein a learning algorithm is tasked with classifying features with a binary ‘true’ or ‘false’ label.

LunchBoxML implements Microsoft’s ML.NET binary classification with a variety of learning algorithm types to select from and calibrate.


Simple Classification Region

This example file uses a Binary Classification Trainer based on random points and their inclusion within a 2D enclosed region.

A Binary Classification tester is used to predict whether a new point is contained within the region.

A color visualization showing the predictive ‘scoring’ of the points from a binary classifier.

Multi-label Classifier using Binary Components

This example shows how users can employ several Binary Classifiers to set up a Multi-Label Classification scenario.

The example trains three binary classification trainers using 3D point collection and their inclusion within three regions.

A new point grid is then tested against the binary classifiers. A point has the possibility being assigned 0-3 labels depending on their inclusion within each of the regions.

Points being tested against binary classifiers for inclusion within up to 3 different regions.
Multiple binary classifiers can be used to establish a multi-label classification scenario.

How can we help?