公开(公告)号：WO2017074770A1

公开(公告)日：2017-05-04

申请号：PCT/US2016/057801

申请日：2016-10-20

Applicant: MICROSOFT TECHNOLOGY LICENSING, LLC

Inventor： TOSAR, Juan Alejandro Llanes ,  SHAMAY, Eric S. ,  ZHENG, Ziying (Eric) ,  YUAN, Flore Q.

IPC: G06F11/36

CPC classification number: G06F11/3692 ,  G06F11/366 ,  G06F11/3684 ,  G06F11/3688

Abstract: Methods, systems, apparatuses, and computer program products are provided for the "bucketing" or categorizing of software failures occurring during software test, and/or during other procedures. Software failure information is received. The software failure information is parsed to generate a raw matrix of software terms, software failures, and an indication of a number of times each software term was found in each software failure. An importance is determined of each software term of the raw matrix with respect to the software failures of the raw matrix to generate a weighted matrix. A failure space is generated based on the determined importance that indicates each software term-software failure pair of the matrix as a vector. A set of clusters representing the vectors in the failure space is determined. Software failures may be automatically assigned to clusters of the set to be "bucketed" or categorized for ease of subsequent analysis.

Abstract translation: 提供方法，系统，装置和计算机程序产品用于“分组” 或在软件测试期间和/或在其他程序期间发生软件故障的分类。 收到软件故障信息。 解析软件故障信息以生成软件术语的原始矩阵，软件故障以及每个软件故障中发现每个软件术语的次数的指示。 确定原始矩阵的每个软件项相对于原始矩阵的软件失败以生成加权矩阵的重要性。 基于所确定的重要性来生成故障空间，该重要性指示矩阵的每个软件术语 - 软件故障对作为向量。 确定表示失败空间中的向量的一组集群。 软件故障可以被自动分配给要被“分组”处理的集合的集群。 或分类以方便后续分析。

公开(公告)号：WO2016183109A1

公开(公告)日：2016-11-17

申请号：PCT/US2016/031689

申请日：2016-05-11

Applicant: MICROSOFT TECHNOLOGY LICENSING, LLC

Inventor： LLANES-TOSAR, Juan A. ,  YUAN, Flore Q. ,  GERSHAFT, Aleks M. ,  ZHENG, Ziying ,  CZERWONKA, Jacek A.

IPC: G06F11/07

CPC classification number: G06N7/005 ,  G06F8/70 ,  G06F11/0706 ,  G06F11/0778 ,  G06F17/16

Abstract: An automated method of making a prediction regarding development of a software product includes receiving code changes information, build information, and failure information related to the software product. Entries are stored in a database, wherein each entry links a subset of the code changes information with a subset of the build information and with a subset of the failure information. A first matrix and a second matrix are generated using the entries in the database. Multi-target entropy calculations are performed based on the first matrix and the second matrix. The prediction regarding the development of the software product is performed based on the multi-target entropy calculations.

Abstract translation: 对软件产品开发进行预测的自动化方法包括接收与软件产品相关的代码更改信息，构建信息和故障信息。 条目存储在数据库中，其中每个条目将代码改变信息的子集与构建信息的子集以及故障信息的子集进行链接。 使用数据库中的条目生成第一矩阵和第二矩阵。 基于第一矩阵和第二矩阵执行多目标熵计算。 基于多目标熵计算进行关于软件产品开发的预测。