This document is a compilation of three separate technical reports. In all detail, the development of a method for detecting the presence of so-called "anti-patterns" in software development projects is presented. The first technical report is concerned with this concrete problem, and it facilitates two major building blocks: The first is the application of a new method for time warping, called self-regularizing boundary time/amplitude warping (srBTAW). The second building block is a detailed walkthrough of creating a classifier for commits, based on source code density. Both these blocks have dedicated technical reports. All of the data, source code, and raw materials can be found online. These reports and resources are made available for reproduction purposes. The interested reader is welcome and enabled to re-run all of the computations and to extend our ideas.
翻译:本文件汇编了三份单独的技术报告,详细介绍了软件开发项目中发现所谓“反模式”存在的方法的开发情况,第一份技术报告涉及这一具体问题,它促进两个主要组成部分:第一个是采用新的时间扭曲方法,称为自我正规化的边界时间/速度扭曲(SrBTAW),第二个构件是根据源代码密度,详细完成创建承诺分类器的工作。这两个构件都有专门的技术报告。所有数据、源代码和原材料都可以在网上找到。这些报告和资源可用于复制目的。感兴趣的读者受到欢迎,能够重新运行所有计算方法,并扩展我们的想法。