基于蜕变关系的贝叶斯分类器测试数据集生成器的设计与实现

Design and Implementation of Bayesian Classifier Test Dataset Generator Based on Metamorphic Relations

蜕变测试方法是一种科学有效的软件测试方法,数据集的生成是人们一直以来的研究重点,本文提出了一种基于蜕变关系的贝叶斯分类器测试数据集生成器。该工具首先将数据集作为输入;其次,依次按照不同的蜕变关系对数据集进行蜕变;最后,输出各个衍生的数据集。该工具实现了针对贝叶斯分类器的测试数据集生成器,提高了贝叶斯分类器测试数据的生成效率,并解决了数据集数量少且质量不高问题。

The metamorphic test method is a scientific and effective software test method.The generation of dataset is the focus of people's research. This paper proposes a Bayesian classifier test dataset generator based on metamorphic relations. The tool first takes dataset as input;secondly,it metamorphic dataset according to different metamorphic relations;finally, outputs the derived dataset.The tool implements a test dataset generator for Bayesian classifiers to improve the efficiency of Bayesian classifier test data generation and solve the problem of a small number of datasets and low quality.