الفهرس 
Data Mining in Testing Large Scale Systems
Test Cases selection TechniquesOnly 14 pages are availabe for public view
 |  | from | 139 |  |  |
| | | from | 139 | | |
Abstract
Regression testing repeatedly executes test cases of previous
builds to validate that any new changes occurred did not affect the
original features. It is the type of software testing that seeks to uncover
new software bugs in existing areas of a system after changes have been
made to them. In recent years, regression testing has seen a remarkable
progress with the increasing popularity of agile methods, which stress on
the central role of regression testing in maintaining the software quality.
The significance of regression testing has grown with the amplified
adoption of agile development methodologies. The optimum case for
regression testing in agile context is to run regression set at the end of
each sprint and release, which requires a lot of cost and time.
In this master’s thesis, we present an automated scalable agile regression
testing approach on both the sprints and release levels. As for the sprints
level, the proposed approach addresses weighted sprint test cases
prioritization technique (WSTP) that prioritizes test cases based on
several agile parameters having real practical weight for testers.
Regarding the release level, two different approaches are proposed:
1. Cluster-based Release Test cases selection technique (CRTS),
which clusters user stories based on the similarity of covered
modules to solve the scalability issue. Test cases are then selected
based on issues logged for failed test cases using text-mining
techniques
2. Regression Testing Reduction and Prioritization (RTRP), which
reduces the number of test cases to be used at regression phasedepending on the similarity of issues exposed from the different
test cases, taking into consideration the user story coverage. It
then prioritizes the reduced test cases using user-provided
weighted agile parameters.
The three different proposed approaches are evaluated using different
evaluation metrics for each technique. The prioritization technique shows
enhancement in the effectiveness of test cases prioritization by average of
APFD equals to 0.78 for the different parameters.
Selection Technique improves the effectiveness of test cases selected by
average of F-measure equals to 0.79 for the different releases, with
different number of word occurrences applied.
Moreover, the reduction and prioritization technique shows an
improvement of TSR by an average of 6%, while retaining the fault
detection capability by an average of 96.5% for the three different
datasets used. As for the prioritization, results show an improvement of
APFD by an average of 0.802 using different weights for the provided
parameters.
In addition, the implemented system execution time is compared to the
manual execution time done by the users. The comparison shows that the
implemented system saves time needed, which consecutively saves the
cost of the regression testing done.