Open AccessPredicting Metamorphic Relations Based on Path Features
Author(s) -
Hanyue Zhang,
Lei Liu,
Peng Zhang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1650/3/032008
Subject(s) - metamorphic rock , computer science , reuse , path (computing) , artificial intelligence , algorithm , construct (python library) , oracle , test (biology) , programming language , engineering , geology , paleontology , geochemistry , waste management
Metamorphic test has been proposed to effectively solve the oracle problem, but the most of the existing metamorphic relations are difficult to reuse, which leads to a large cost of in the metamorphic test. In order to improve the efficiency of the metamorphic test and solve the problem of low reuse rate of metamorphic relations, based on the common set of metamorphic relations in scientific computing programs and execution path, we propose a novel string feature with a new extraction method. Then, this feature can be used to train support vector machine to decide the metamorphic relations for test. At last, we construct a various of experiments, and the experiments results show that our method can effectively predict the satisfaction of the input features for the metamorphic relations.