Evolving Similarity Functions for Graph-based Semi-Supervised Learning

doi:10.1109/tevc.2026.3717536

Evolving Similarity Functions for Graph-based Semi-Supervised Learning

2026-07-27

0

OA

AI

DOI：10.1109/tevc.2026.3717536

Original

Original request for help

Save

Abstract

En 中文

Automatic data classification is commonly performed using supervised learning algorithms that build models to classify new instances. However, labelled data are often scarce and expensive to obtain. Semi-supervised learning (SSL) offers a viable alternative by exploiting a few labelled and many unlabelled instances. Among various SSL approaches, graph-based methods are particularly appealing as they rely on neighbourhood structure rather than explicit parametric distributional assumptions. Nevertheless, their performance heavily depends on graph construction, which requires defining a suitable similarity (or distance) function. Since no single similarity function performs optimally across all datasets, this paper introduces two evolutionary approaches – GEAS and ST-GPAS – to automatically design similarity functions for graph-based SSL. Extensive experiments on 50 public datasets show that the proposed methods yield similarity functions leading to superior transductive learning performance compared with traditional distance-based measures and representative metric-learning algorithms. The improvements, measured in Micro-F₁ and Macro-F₁, are statistically significant under nonparametric tests. GEAS and ST-GPAS achieve closely matched performance, with small but consistent advantages depending on the supervision regime.

Keywords:

Transductive semi-supervised learning

graph-based algorithms

similarity functions

grammatical evolution

strongly-typed genetic programming

Evolving Similarity Functions for Graph-based Semi-Supervised Learning

Abstract

Journal

Organization

Cited Papers