A multi-instance learning algorithm based on a stacked ensemble of lazy learners

TL;DR

This paper introduces a multi-instance learning algorithm that uses a stacked ensemble of lazy learners, optimized for accuracy and false positive rate, demonstrating effectiveness on the Musk1 dataset.

Contribution

It presents a novel ensemble-based multi-instance learning algorithm using lazy learners and multi-objective optimization for parameter diversity.

Findings

Effective on Musk1 benchmark dataset

Optimizes for accuracy and false positive rate

Ensemble improves classification performance

Abstract

This document describes a novel learning algorithm that classifies "bags" of instances rather than individual instances. A bag is labeled positive if it contains at least one positive instance (which may or may not be specifically identified), and negative otherwise. This class of problems is known as multi-instance learning problems, and is useful in situations where the class label at an instance level may be unavailable or imprecise or difficult to obtain, or in situations where the problem is naturally posed as one of classifying instance groups. The algorithm described here is an ensemble-based method, wherein the members of the ensemble are lazy learning classifiers learnt using the Citation Nearest Neighbour method. Diversity among the ensemble members is achieved by optimizing their parameters using a multi-objective optimization method, with the objectives being to maximize Class 1 accuracy and minimize false positive rate. The method has been found to be effective on the Musk1 benchmark dataset.