International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 7 (2018) pp. 5167-5174
© Research India Publications. http://www.ripublication.com
5167
Test Case Retrieval Model by Genetic Algorithm with UML Diagram
Itti Hooda
(1),
, Prof. Dr. Rajender Singh Chhillar
(2)
1
Department of Computer Science and Applications, Maharishi Dayanand University, Rohtak,(Haryana), India.
2
Department of Computer Science and Applications, Maharishi Dayanand University, Rohtak,(Haryana), India.
Abstract
In this research paper we have shown the Reusability of test
cases. If companies make repositories of Test cases then they
can automatically extract the cases from Database, but the
problem is matching of the test cases and the new Project
UML Design. So we have used optimize weight model by
Genetic algorithm which match the test cases by cosine
similarity and extract the High Cosine similarity test cases
which analysis on the basis of TPR and FPR (True positive
rate and False positive rate). We can also find the accuracy of
proposed process. We achieve significance high accuracy like
in case of ATM 83%, Student management System 81.23%
and Banking System 80.23%.
Keywords: FPR, IF Methods, LDA, BLUiR,UML,
Mutation,Crossover,GA.
INTRODUCTION
Modern society depends on software-intensive systems.
Software operates intangibly or very quietly in everything
from kitchen appliances to critical infrastructure, and one's
living a life exclusive of every day relying on systems running
software needs a determined downshifting from life as most
people enjoy it. As the importance of software constantly
grows, so does the significance of being able to create it
proficiently. Software development is an wide-ranging
expression used to illustrate any approach to produce source
code and its correlated documentation. Through the software
crisis of the 1960s, it became clear that software difficulty
quickly rises when software are scaled up to larger systems.
The development techniques that were applied at the time did
not result in needed software in a predictable manner. [7]
A powerful search engine that can retrieve software artifacts
with high precision is obviously of great importance to
developers. Software systems usually contain defects that
need to be fixed after releases, and in some projects users are
allowed to submit feedback on these defects that they
encounter through bug reporting systems such as Bugzilla.
Owing to this inherent complexity of software construction,
software bugs remain frequent. For a large software system,
the number of bugs may range from hundreds to thousands.
However, performing this process manually for lot of bugs is
time consuming and exclusive. Consequently, efficient
techniques for locating bugs automatically from bug reports
are highly desirable [1].In recent years, information retrieval
(IR) based bug localization techniques have gained significant
attention owing to their relatively low computational cost and
minimal external dependencies. Effectively supporting
software changes is essential to give a sustainable high-quality
evolution of large-scale software systems, as realizing even a
slight change may not be always straightforward. Software-
change impact analysis, or basically impact analysis (IA), has
been acknowledged as one such key protection activity. IA
aims at estimating the potentially impacted entities of a
system due to a proposed change [5].
The goal of contextual search is to incorporate a richer model
of human searching behaviour into search systems, and as
such, represents an opportunity for collaboration between
information behaviour (IB) and information retrieval (IR), two
fields which have hitherto progressed mostly in parallel.
Researchers have worked towards maximizing the impact that
software engineering research has on practice, for example, by
providing methods and outputs that are as general (and thus as
useful) as possible. A significant amount of research on
applying Information Retrieval (IR) methods for analysing
textual information in software artifacts has been conducted in
the SE community in recent years. Software testing is
indispensable for all software development. It is an integral
part of the software engineering discipline. However, testing
is labour-intensive and exclusive. It often accounts for more
than 50% of total development costs. Thus, there are clear
advantage in reducing the cost and improving the efficiency of
software testing by automating the process. [9].
In fact, there has been a quick expansion of practices in using
automated software testing tools. Currently, a large number of
software test automation tools have been developed and have
become accessible on the market. Between a range of testing
activities, test case invention is one of the most intellectually
demanding tasks and it is also of the most critical challenges,
since it can have a tough impact on the effectiveness and
efficiency of the whole testing process The goal of
performance testing is to find performance problems, when an
application under test (AUT) unexpectedly exhibits worsened
characteristics for a specific workload. Existing feature
location techniques use different tactics to find a feature’s
source code. Approaches based on information retrieval (IR)
leverage the fact that identifiers knowledge to locate source
code that is textually similar to a query describing a feature is
there [4].
Various researchers have applied information retrieval
methods to automatically search for appropriate files based on
bug reports. They treat an initial bug report as a query and
rank the source code files by their significance to the query.
The developers can then analysis the returned files and fix the
bug. These techniques are information retrieval based bug
localization methods. Numerous of the obtainable IR-based