Automatic evaluation method for Android compatibility defect repair effect

Patent

Automatic evaluation method for Android compatibility defect repair effect

- 20 Nov 2020

TL;DR: In this article, a method for automatically evaluating the repairing effect of Android compatibility defects using static analysis and machine learning has been proposed, which comprises the specific steps of code compatibility defect repair static analysis, feature extraction and effect evaluation.

Abstract: The invention belongs to the technical field of Android system compatibility testing, and particularly relates to a method for automatically evaluating the repairing effect of Android compatibility defects. The compatibility defect repair schemes comprise two schemes, namely simple compatibility defect inspection and replaceable compatibility defect inspection; according to the method, static analysis and machine learning are utilized to evaluate the two compatibility defect repair schemes; the method comprises the specific steps of code compatibility defect repair static analysis, compatibility defect repair feature extraction and compatibility defect repair effect evaluation, namely training and learning extracted feature vectors to obtain a final model; the model can evaluate the code compatibility check of an unknown application and output a corresponding compatibility repair scheme. According to the method, the repairing mode of API compatibility defects in the application programcan be automatically identified, and the repairing effect is classified, so that beneficial help is provided for Android platform maintainers and developers.

Chat with Paper

AI Agents for this Paper

Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps

References

Proceedings Article•10.1145/3377811.3380357

How Android developers handle evolution-induced API compatibility issues: a large-scale study

Hao Xia, +10 more

- 27 Jun 2020

TL;DR: RAPID, an automated tool to determine whether a compatibility issue has been addressed or not, by incorporating both static analysis and machine learning techniques is proposed, which would be extremely useful to novice developers and may significantly improve the current status of compatibility issue handling.

...read moreread less

46

Patent

Python program type defect detection method based on abstract syntax trees

Chen Lin, +3 more

- 04 Dec 2018

TL;DR: In this article, a Python program type defect detection method based on abstract syntax trees is proposed, which comprises the steps of collecting Python software defect report information and extracting a defect identifier and a defect error type, obtaining source codes of two versions of programs before and after defect repair, generating abstract syntax tree corresponding to the twoversions of source codes, matching and obtaining a change function node and marking the defect error types, generating a feature vector of a defect code according to the context information of the change function nodes, training a multi-classification models on a defect feature vector by using

...read moreread less

9

Patent

Self-repair system and method for providing resource failure tolerance

Erik H. Volkerink, +1 more

- 17 Feb 2006

TL;DR: In this article, a self-repair system provides resource failure tolerance using an interconnection network that provides interconnection information identifying connections between system resources, redundant resources and ports that are connectable to consumers of the system resources.

...read moreread less

8

Patent

Static analysis and neural network based software failure prediction technique construction method

Shunkun Yang, +11 more

- 23 Mar 2018

TL;DR: In this paper, a static analysis and neural network based software failure prediction technique construction method is presented, which includes the following steps: 1, collecting valid failures of diagnosed software, adding the valid failures into a created failure case library, counting the frequency of historical versions of valid failures, and outputting a complexity metric; 2, performing correlation analysis, and calculating the significance level of the failure frequency and the metric; 3, constructing a network training input output matrix and a prediction input matrix; 4, completing network training, and 9, performing neural network prediction, and 10, performing

...read moreread less

3

Patent

Static analysis method and tool based on defect mode

Liu Lei, +4 more

- 04 Jul 2017

TL;DR: In this article, a static analysis method and tool based on a defect mode is presented, which comprises the steps that the defect mode of a program is pre-edited; files are traversed, and the file needing analysis is found; a source program is scanned line-by-line to remove codes not needing analysis; and defect analysis is performed, the defectmode is called to be matched with the program without the codes, and if matching succeeds, it is indicated that the file has a defect.

...read moreread less

2