Unsupervised Classification by Iterative Voting

Abstract

In the paper we present a simple algorithm for unsupervised classification of given items by a group of agents. The purpose of the algorithm is to provide fast and computationally light solutions of classification tasks by the randomly chosen agents. The algorithm follows basic techniques of plurality voting and combinatorial stable matching and does not use additional assumptions or information about the levels of the agents' expertise. Performance of the suggested algorithm is illustrated by its application to simulated and real-world datasets, and it was demonstrated that the algorithm provides close to correct classifications. The obtained solutions can be used both separately and as initial classifications in more complicated algorithms.

Distinguishing the Leading Agents in Classification Problems Using the Entropy-Based Metric

“…To avoid such influence, in more sophisticated methods [10][11][12][13][14], the problem is divided into two stages. First, using the agents' classifications γ k , k = 1, 2, .…”

“…Another approach to selecting the competent agents was implemented in the algorithms [12,13], which are based on the similarities of the agents' classifications, and in the algorithm [14], where the competent agents are selected using the expectation bias [15].…”