首页> 外国专利> SOURCE CODE EVALUATION DEVICE, SOURCE CODE EVALUATION METHOD AND SOURCE CODE EVALUATION PROGRAM

SOURCE CODE EVALUATION DEVICE, SOURCE CODE EVALUATION METHOD AND SOURCE CODE EVALUATION PROGRAM

机译:源代码评估设备,源代码评估方法和源代码评估程序

摘要

To provide a source code evaluation device capable of automatically estimating a location where a source is likely to contain a bug.SOLUTION: In the source code evaluation device 10, a difference detection unit 11 detects a description to be a difference in source code before and after the source code is changed, and determines whether the difference is originated from a correction of a bug. A lexical-vector generating unit 12 divides the description to be a difference into lexical tokens and generates a vector in which numerical values defined according to a type of each lexical token are arranged in order of individual lexical tokens in the description to be a difference. A learning unit 13 learns a relationship between the vector and a result of determination by the difference detection unit. A window setting unit 14 sets a window containing a predetermined number of lexical tokens for a source code to be evaluated for presence/absence of a bug. An evaluation target analysis unit 15 generates a vector for each lexical token contained in the window. Based on the generated vector and a result of the learning, a recognition unit 16 evaluates a likelihood of a bug being contained in a description within the window.SELECTED DRAWING: Figure 2
机译:提供一种能够自动估计源中可能包含漏洞的位置的源代码评估装置。解决方案:在源代码评估装置10中,差异检测单元11将描述描述为源代码之前和之后的差异。更改源代码后,确定差异是否源自错误的更正。词汇向量生成单元12将作为差异的描述划分为词汇标记,并生成向量,其中根据每个词汇标记的类型定义的数值以描述中的各个词汇标记的顺序排列为差异。学习单元13学习向量与差异检测单元的确定结果之间的关系。窗口设置单元14设置包含预定数量的词法标记的窗口,该词法标记用于要评估是否存在错误的源代码。评估目标分析单元15为窗口中包含的每个词汇标记生成向量。基于生成的向量和学习结果,识别单元16评估窗口中的描述中包含错误的可能性。

著录项

  • 公开/公告号JP2019021037A

    专利类型

  • 公开/公告日2019-02-07

    原文格式PDF

  • 申请/专利权人 NIPPON TELEGR & TELEPH CORP NTT;

    申请/专利号JP20170139091

  • 发明设计人 TABATA KEIICHI;HOSHINO TAKASHI;

    申请日2017-07-18

  • 分类号G06F11/36;G06F8/70;

  • 国家 JP

  • 入库时间 2022-08-21 12:20:08

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号