Symmetric bug prediction in software requirement by machine learning algorithms

利用机器学习算法对软件需求中的对称缺陷进行预测

阅读:1

Abstract

In the software development, high impact on diverse industries are expanding at a fast pace. Early defect fixation and adjustment according to needs are essential to increase the performance, accuracy, durability and reliability of the software. Machine Learning (ML) approaches have used to analyze the Software Requirement (SR) in software development offer a great way to accelerate symmetric development and defect correction, especially because software failures are unpredictable. This research utilizes classification and regression-based models like Random Forest (RF), Artificial Neural Network (ANN), and Adaptive Moment Estimation (AME) to forecast bug resolving times from applicable asymmetric data attributes. ML models are used to predict defect resolution and feature completion time. The data takes into consideration several stages including issue detection, testing, and validation. The outcome of ML models, K-Nearest Neighbors (KNN) algorithm has an accuracy of 66% but the proposed approach of RF, ANN, and ADE illustrates an astonishing 98% accuracy. The dataset with symmetric attributes has trends and strong correlations. The proposed model result is better than other available software failure time prediction in terms of precision and accuracy, providing a sound to defect solution and prediction in SR.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。