Using Semi-Supervised Learning for Predicting Metamorphic Relations

TL;DR

This paper presents a semi-supervised learning approach to predict metamorphic relations in software testing, reducing the need for expert input and improving prediction accuracy compared to supervised methods.

Contribution

It introduces a semi-supervised machine learning method for predicting metamorphic relations, enhancing accuracy by leveraging unlabeled data.

Findings

Semi-supervised model outperforms supervised model in accuracy.

Unlabeled data significantly improves metamorphic relation prediction.

Method reduces reliance on domain experts for test case generation.

Abstract

Software testing is difficult to automate, especially in programs which have no oracle, or method of determining which output is correct. Metamorphic testing is a solution this problem. Metamorphic testing uses metamorphic relations to define test cases and expected outputs. A large amount of time is needed for a domain expert to determine which metamorphic relations can be used to test a given program. Metamorphic relation prediction removes this need for such an expert. We propose a method using semi-supervised machine learning to detect which metamorphic relations are applicable to a given code base. We compare this semi-supervised model with a supervised model, and show that the addition of unlabeled data improves the classification accuracy of the MR prediction model.