Earticle

Software maintenance is one of the major activities of the software development life cycle. Due to the time and cost constraint it is not possible to perform exhaustive regression testing. Thus, there is a need for a technique that selects and prioritizes the effective and important test cases so that the testing effort is reduced. In an analogous study we have proposed a new variable based algorithm that works on variables using the hybrid technique. However, in the real world the programs consist of multiple modules. Hence, in this work we propose a regression testing algorithm that works on interprocedural programs. In order to validate and analyze this technique we have used various programs. The result shows that the performance and accuracy of this technique is very high.

소프트웨어 패치가 빈번하게 이루어지는 최근의 추세에 따라, 소프트웨어 버그 역시 패치로 인해 유도되는 버그인 회귀 버그의 비중이 점차 증가하는 추세이다. 이에 산업계와 학계에서는 최근 자동 버그 탐지 방법으로 주목받고 있는 퍼즈 테스팅을 도입 및 개량하여 회귀 버그를 사전에 탐지하고자 하는 시도가 점차 활발해지고 있다. 이 논문에서는 회귀 버그 탐지를 위한 퍼즈 테스팅 연구의 현황에 대하여 살펴보고, 현재 기법들에 존재하는 한계를 참고삼아 향후 관련 연구의 방향에 대한 전망을 제시한다.

As software gets an increasing amount of patches, lots of software bugs are increasingly caused by such software patches, collectively known as regression bugs. To proactively detect the regressions bugs, both industry and academia are actively searching for a way to augment fuzz testing, one of the most popular automatic bug detection techniques. In this paper, we investigate the status quo of the studies on augmenting fuzz testing for regression bug detection and, based on the limitations of current proposals, provide an outlook of the relevant research.

행동과학 이론의 질적 성장에 따라 연구자들은 예측변수와 준거변수 간의 복잡한 관계를 이해하기 위해 매개와 조절 모형의 통합에 높은 관심을 가지게 되었다. 대표적인 연구방법이 매개된 조절과 조절된 매개이다. 매개된 조절은 조절효과가 “왜” 또는 “어떻게” 발생하는지를 제 3의 변수인 매개변수를 통해 밝히고자하며 통계적으로 매개효과 검정에 해당된다. 그리고 조절된 매개는 매개과정(mediation process)이 조절변수의 수준에 따라 달라지는 경우에 발생하며 통계적으로 조건부 간접효과를 검정한다. Hayes의 PROCESS는 경로분석 접근법에 따른 조건부 과정분석(conditional process analysis)을 위해 개발된 SPSS 매크로 프로그램이다. PROCESS는 다양한 통합 모형에서 조건부 간접효과를 용이하게 검정할 수 있다는 장점이 있다. 이 프로그램의 성장과 맞물려 조절된 매개효과 모형은 빈번하게 활용되고 있다. 이에 반해 매개된 조절효과 검정은 상대적으로 그렇지 못하다. 매개된 조절의 활용이 저조한 대표적인 이유는 많은 연구자들이 매개된 조절은 조절된 매개와 통계적으로 구분되지 않고 PROCESS를 활용한 자료 분석 절차도 동일하다고 믿기 때문이다. 이에 본 논문은 조절회귀분석(moderated multiple regression)에 기반을 둔 세 가지 유형(TypeⅠ, Ⅱ, Ⅲ)의 매개된 조절효과 모형을 소개한다. 매개된 조절을 최초로 제안한 Baron과 Kenny(1986)의 단계적 접근법(causal steps approach)을 활용하여 이들 유형에 대해 각각의 통계적 정의 및 자료 분석 절차를 논의한다. 먼저, 이들의 탄생 배경을 이해하기 위해 조절회귀모형에서 조절효과의 개념적 모형(conceptual model)과 통계적 모형(statistical model)의 차이를 논의한다. 다음으로 각 유형 별로 단계적 접근법에 따른 매개와 조절 모형의 통합과정, 통계적 모형, 매개된 조절효과 검정 절차를 논의한다. 마지막으로, 이러한 논의를 바탕으로 PROCESS(Version 3.1)를 활용한 Type III 매개된 조절효과 검정 절차를 예시한다. 연구자들이 연구 배경과 연구 문제에 따라 적합한 조절효과 모형과 매개된 조절효과 유형을 선택하고 타당한 절차에 따라 검정을 진행하는데 도움이 되고자 한다.

Baron and Kenny (1986) first proposed the term mediated moderation, to explore how and why the moderation effect occurs. Most researchers believe that Baron and Kenny (1986)’s model is the only mediated moderation method for analysis. In fact, there are three types of mediated moderation models (Type I, II and III) that have each been adopted as most appropriate in various research areas of behavioral science. However, few are aware of existence of these three types and their distinctive conceptual and statistical definitions. Additionally, to date there has been no source that provides an integrated introduction of the methodological, statistical analyses to three types of mediated moderation model. It is our intention in this manuscript to provide a comprehensive, accessible framework by which researchers examining mediated moderation models can effectively choose a valid conceptual model and statistical testing procedures. We first, define each of three types of the mediated moderation model based on the differences between their conceptual and statistical models and provide examples of each. Second, we discuss how the three types were derived from a single concept of mediated moderation model via the approach of Baron and Kenny (1986). Third, we describe the analytic procedure by examining each mediated moderation model based on a statistical significance testing approach with a real example, accompanied by a regression-based software, PROCESS (version 3.1).

본 연구는 낙동강 대권역에서 공간지리 가중회귀모형을 이용하여 녹지지역과 BOD, TN, TP를 포함하는 수질과의 지역적 비균질적 관계를 검증하고자 수행되었다. 대부분 기존의 상관분석 혹은 회귀분석은 OLS (Ordinary Least Square)기법에 기초한 균질적 관계의 분석에 초점을 두어 왔다. 이러한 녹지지역과 수질의 균질적 관계에 기초한 분석은 지역적으로 변화를 고려하지 않는 단점이 있다. 연구대상지는 낙동강 대권역내 146개 지점이며, 수질자료와 토지피복 자료는 환경부 자료를 활용하였다. BOD, TN, TP에 대하여 일반 회귀모형과 모델과 GWR 모델을 추정하여 비교하였다. 비교결과, BOD와 TN의 GWR 모델이 OLS 모델에 비하여 우수한 것으로 나타났다. GWR 모델의 $R^2$와 녹지지역의 계수 값의 기초통계량을 분석한 결과, 지역적으로 큰 변동이 확인되었다. 이러한 결과는 유역 토지이용과 수질과의 관계가 공간적으로 비균절적이라는 것을 입증하여 주었다.

This study aims to examine the presence of non-stationary relationship between water quality and land use in watersheds. In investigating the relationships between land use and water quality, most previous studies adopted OLS method which is assumed stationarity. However, this approach is difficult to capture the local variation of the relationships. We used 146 sampling data and land cover data of Korean Ministry of Environment to build conventional regressions and GWR models for BOD, TN and TP. Regression model and GWR models of BOD, TN, TP were compared with $R^2$, AICc and Moran's I. The results of comparisons and descriptive statistics of GWR models strongly indicated the presence of Non-Stationarity between water quality and land use.

The process of verifying the modified software in the maintenance phase is called Regression Testing. The size of the regression test suite and its selection process is a complex task for regression testers because of time and budget constraints. In this research paper, the Bee Colony Optimization (BCO) algorithm for the fault coverage regression test suite prioritization has been presented. In the natural bee colony, there are of two types of worker bees; Scout bees and forager bee, who are responsible for the development and maintenance of the colony. The BCO algorithm developed for the fault coverage regression test suite is based on the behavior of these two bees. The BCO algorithm has been formulated for fault coverage to attain maximum fault coverage in minimal units of execution time of each test case, using two examples whose results are comparable to optimal solution. Average Percentage of Fault Detection (APFD) metrics and charts has been used to show the effectiveness of proposed algorithm.

The delivery of new increments involves decision makings at various levels like decision aspect prioritization, requirement priorities and selections of regression test cases. The pressure to deliver high quality low cost software by employing limited resources and satisfying mass market invokes the need to have less dependence on regression testing activity. Earlier Proposed requirement prioritization technique categorizes the type of changes and the type of new implementations and then decides the neglection or minimization of test cases. Further minimizations are also possible due to existence of sound regression testing techniques. By focussing on common problems related to the development of mobile and embedded type software applications, one could have great improvement in software development quality and practices. The proposed requirement prioritization technique is applied in evolutionary mobile software application development and results were quite promising. This paper aims to have a deep insight into the ability of the proposed technique in minimization of regression testing of desktop software applications application. The results are very promising thus indicating the usefulness of proposal in low cost development of both mobile compatible and desktop software applications. The novel contribution of this paper lies in making the incremental development timely and low cost activity by making requirement prioritization effective and future increments less dependent on regression testing without having any compromise with the desktop software quality. This had been a result of analysis of proposed prioritization technique on desktop software application.

We propose, in this paper, a hybrid regression testing technique and associated tool for object-oriented software. The technique combines, in fact, the analysis of UML models to a simple static analysis of the source code of the modified program. The basic models we use are use cases model and corresponding UML statechart and collaboration diagrams. The goal of the static analysis of the source code is to identify changes that are not visible in design models. The developed tool identifies the modified (and/or impacted by modifications) use cases and selects the appropriate test cases from an existing test suite. New (JUnit) test cases, covering new scenarios or those whose structure has been modified after changes, are generated when necessary. In this way, the technique supports an incremental update of the test suite. The selected JUnit test cases, including the new ones, are automatically executed. A case study is reported to provide evidence of the feasibility of the approach and its benefits in terms of reduction of regression testing effort.

This research discusses the problems about scope accuracy and quality control in application of regression testing in the engineering practice, and proposes a practical regression method, combing with change-impact-analysis, business rules model, cost risk assessment and test case management. This approach has already been applied to functionality testing of some core systems in our domestic financial institutions, and made great achievement.

Due to dynamic changing business and operating environment, software evolution is inevitable. As software evolves, there is a need for a retest of the system in order to ensure its validity and thus regression testing assures that this happens. Regression Testing has several research categories, one of the most popular areas is Test Case Prioritization which orders test cases based on a particular criteria to be tested in accordance to available resources, and thus ensuring that the most critical test cases are tested first. In this study we examine the different types of test case prioritization and a description of the techniques presented alongside the issue raised in each study. This will give an insight into the main problems currently plaguing test case prioritization and thus pave way for further research areas to tap into.

In this article, we describe the regression test process to test and verify the changes made on software. A developed technique use the automation test based on decision tree and test selection process in order to reduce the testing cost is given. The developed technique is applied to a practical case and the result show its improvement.

After the version changes,locatingthe influence domain of control flow modificationis a hot issue in software testing, because it helps developersto solve software defects effectivelyand testersto determine the regression test cases. This paper puts forward that an algorithm toanalyzethe influence domain of control flow modification. It can accurately locate the functions which are affected by modified control statements. According to the control flow function calling paths algorithm, a path of nodes is extracted, where control word as well as the function name is regarded as a node,and extract the control block and analysis it, and a static function calling path with the control logic is obtained.Control flow modificationpoint algorithm is put forward to obtain the modified control statements, which is based on Longest Common Substring (LCS) algorithm. An algorithm is also proposed to determine the influence domain of control flow modification.Experimental results show that the proposedalgorithm can determine the influence domain, and provide support for test case reduction in regression testing effectively.

Some tests of the goodness of fit of the uniform distribution between 0 and 1 are presented. The powers of the tests under certain alternatives are examined. As a result, the statistic based on the difference between the order statistics and the modal value of them gives good powers. We also give modifications of the statistic without using the extensive tables of the critical points.

Given the specific mean shift outlier model, several standard approaches to obtaining test statistic for outliers are discussed. Each of these is developed in detail for the nonlinear regression model, and each leads to an equivalent distribution. The geometric interpretations of the statistics and accuracy of linear approximation are also presented.

회귀분석은 반응변수와 예측변수들 간의 관련성을 설명하기 위해 사용되는 잘 알려진 통계 테크닉이다. 특히 연구자들은 두 개의 독립 모집단에서의 모형들의 회귀계수들(절편과 기울기)을 비교하는데 관심이 있다. Gregory Chow에 의해 제안된 Chow 검정은 회귀모형들을 비교하고 선형회귀모형 안에 구조적 브레이크가 존재하는지를 검정하기 위해 보통 사용되는 방법들 중의 하나이다. 본 연구에서는 두 독립 선형회귀모형들의 등가성을 검정하기 위해 퍼뮤테이션 방법을 제안하고 Chow 검정과 비교한다. 그리고 퍼뮤테이션 검정과 Chow 검정의 검정력을 조사하기 위해 시물레이션 연구를 진행하였다.

Regression analysis is a well-known statistical technique useful to explain the relationship between response variable and predictor variables. In particular, Researchers are interested in comparing the regression coefficients(intercepts and slopes) of the models in two independent populations. The Chow test, proposed by Gregory Chow, is one of the most commonly used methods for comparing regression models and for testing the presence of a structural break in linear models. In this study, we propose the use of permutation method and compare it with Chow test analysis for testing the equality of two independent linear regression models. Then simulation study is conducted to examine the powers of permutation test and Chow test.

This paper consider the testing problem of grouped heteroscedasticity in the linear regression model. We provide the Lagrange Multiplier(LM), Wald, Likelihood Ratio (LR) test statistis for testing of grouped heteroscedasticity. Monte Carlo experiments are conducted to study the performance of these tests.

Consider a linear model with n observations and k explanatory variables: (1)b $y=X\beta+u, u\simN(0,\sigma^2I_n)$. We assume that the model satisfies the ideal conditions. Consider the general linear constraints on regression coefficient vector: (2) $R\beta=r$, where R and r are known matrices of orders $q\timesk$ and q\times1$ respectively, and the rank of R is $q<k$. We also assume $n>k+q$.

In this paper we consider the problem of identifying and testing the outliers in linear regression. first we consider the problem for testing the null hypothesis of no outliers. The test based on the ratio of two scale estimates is proposed. We show the asymptotic distribution of the test statistic by Monte Carlo simulation and investigate its properties. Next we consider the problem of identifying the outliers. A forward sequential procedure based on the suggested test is proposed and shown to perform fairly well. The forward sequential procedure is unaffected by masking and swamping effects because the test statistic is based on robust estimate.

Several information criterions, Schwarz information criterion (SIC), Akaike information criterion (AIC), and the modified Akaike information criterion ($AIC_c$), are proposed to locate a change point in the multiple linear regression model. These methods are applied to a stock Exchange data set and compared to the results.

In econometric field, a question often arises whether the consumption pattern at a later time is the same as the one found before. This question is statistically answered by testing whether an additional observation belongs to a regression estimated by a group of observations obtained before. A diagnostic method of investigating the influence of observations in this testing problem is suggested. A perturbed test statistic for an appropriately defined perturbation scheme is computed, and then this and the perturbation vector form a surface in the Euclidean space as the perturbation vector varies over a certain region. The maximum direction vector associated with the largest curvature of the curve on this perturbation-formed surface at the null point provides information about influential observations for this testing problem. Some necessary computations are provided for finding curvatures and direction vectors. A numerical example is given for illustration.

We consider the problem of identifying and testing outliers in linear regression. First, we consider the scale-ratio tests for testing the null hypothesis of no outliers. A test based on the ratio of two residual scale estimates is proposed. We show the asymptotic distribution of test statistics and investigate the properties of the test. Next we consider the problem of identifying the outliers. A forward procedure based on the suggested test is proposed and shown to perform fairly well. The forward procedure is unaffected by masking and swamping effects because the test statistics used a robust scale estimate.