基于LDA主题模型的雷达软件缺陷分类算法研究

Research on radar software defect classification algorithm based on LDA topic model

针对测试验证时雷达软件缺陷预测精度和效率不高的问题,提出一种基于LDA主题模型的雷达软件缺陷分类算法,融入面向雷达领域的分词字典,实现对雷达软件的准确分词。通过获取以功能、性能和接口等需求为导向的隐含主题,有效提高雷达软件缺陷分类的准确度和有效性,进而提升雷达软件测试设计工作的质量和效率。

Due to the low accuracy and efficiency of software defect prediction during the testing and verification,a radar software defect classification algorithm is proposed based on the Latent Dirichlet Allocation(LDA)topic model,which integrates the radar-oriented word segmentation dictionary to achieve accurate word segmentation of radar software.The accuracy and effectiveness of radar software defect classification can be effectively improved by acquiring implicit themes oriented towards requirements of function,performance and interface,thereby improving the quality and efficiency of radar software testing and design.