Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/461465
Title: | Software fault prediction techniques in software testing |
Researcher: | Mohapatra, Yogomaya |
Guide(s): | Ray, Mitrabinda |
Keywords: | Computer Science Computer Science Software Engineering Engineering and Technology |
University: | Siksha |
Completed Date: | 2021 |
Abstract: | In software development, dealing with software faults is vital and important task. One of the key challenges of the tester is to find the test suite quality, which is to identify defects. The faults reduce the software quality and increase its development cost. In traditional fault prediction techniques, previous knowledge of faults or a faulty module is required while detecting software faults in an application. In this thesis, various defect prediction models using Machine Learning (ML) techniques are proposed to identify faulty modules. These models help to enhance the quality of software by reducing the faults, execution time and testing cost. The software fault prediction (SFP) model trains the learning techniques to produce base learners which are then applied to unknown projects. Artificial Neural Network (ANN) has been widely used in fault prediction. Average Percentage of Faults Detected (APFD) and Problem Tracking Report (PTR) are the two familiar software metrics used for measuring the test efficiency of the generated test suite. Work has been done on fault prediction using ANN; however it is observed that APFD is low using only ANN in the fault prediction software model. There is an optimization algorithm, called Cuckoo Search (CS) which resembles the breeding activity of cuckoos. It searches for better solution by replacing the worst solution in the nest. A hybrid ANN-CS fault prediction model is proposed, where CS is used for optimizing the weight factor of ANN. The Cyclomatic Complexity metric of the case study, Hospital Management System (HMS), for which the fault has to be identified, is input to the proposed model. Here, the test cases are generated automatically from the source code using Code Pro tool. The test case features like execution time, failure, line coverage and loop coverage are extracted. Then, test cases with these features are clustered by the help of k-means clustering. After clustering, the test cases are input to the modified ANN where the faulty and faultless classes are clas |
Pagination: | |
URI: | http://hdl.handle.net/10603/461465 |
Appears in Departments: | Department of Computer Science |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 421.57 kB | Adobe PDF | View/Open |
02_prelim pages.pdf | 1.02 MB | Adobe PDF | View/Open | |
03_content.pdf | 262.65 kB | Adobe PDF | View/Open | |
04_abstract.pdf | 151.57 kB | Adobe PDF | View/Open | |
05_chapter 1.pdf | 536.94 kB | Adobe PDF | View/Open | |
06_chapter 2.pdf | 500.42 kB | Adobe PDF | View/Open | |
07_chapter 3.pdf | 1.01 MB | Adobe PDF | View/Open | |
08_chapter 4.pdf | 1.08 MB | Adobe PDF | View/Open | |
09_chapter 5.pdf | 926.6 kB | Adobe PDF | View/Open | |
10_chapter 6.pdf | 349.21 kB | Adobe PDF | View/Open | |
11_annexures.pdf | 508.39 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 769.82 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: