-
[in Japanese]
2019 Volume 36 Issue 4 Pages
4_1
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Akito MONDEN, Kei ITO, Toshihiro KAMIYA, Hirohisa AMAN, Noriko HANAKAW ...
2019 Volume 36 Issue 4 Pages
4_2
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Hiroyuki KIRINUKI, Haruto TANNO, Katsuyuki NATSUKAWA
2019 Volume 36 Issue 4 Pages
4_3-4_17
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
Test automation tools such as Selenium are commonly used for automating end-to-end tests, but when developers update the software, they often need to modify the test scripts accordingly. However, the costs of modifying these test scripts are a big obstacle to test automation because of the scripts' fragility. In particular, locators in test scripts are prone to change. Some prior methods tried to repair broken locators by using structural clues, but these approaches usually cannot handle radical changes to page layouts.
In this paper, we propose a novel approach called COLOR (correct locator recommender) to support repairing broken locators in accordance with software updates. COLOR uses various properties as clues obtained from screens (i.e., attributes, texts, images, and positions). We examined which properties are reliable for recommending locators by examining changes between two release versions of software, and the reliability is adopted as the weight of a property. Our experimental results obtained from four open source web applications show that COLOR can present the correct locator in first place with a 77% – 93% accuracy and is more robust against page layout changes than structure-based approaches.
View full abstract
-
Yoshinari HACHISU, Atsushi YOSHIDA, Hiroaki KUWABARA, Kiyoshi AGUSA
2019 Volume 36 Issue 4 Pages
4_18-4_24
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
We propose a method for generating input forms on a Web-based C programming learning environment from input function calls in a source program. While Web programming environments have been becoming popular, their input on executing a program is not suited for Web, because it is based on command line input. We have proposed association array model for Web input, instead of stream model. We design new input functions instead of scanf function, which is usually used on learning the C language. Our Web programming environment analyzes a C source program before run, and generates Web forms dynamically. Names of forms are keys and input data are values of an association array. When a form has multiple values, they are treated as a list with iterator functions. We confirmed that we can compile and run sample programs of a textbook using our input functions and Web forms.
View full abstract
-
Kazuyuki HIGASHI, Hitoshi TAKAHASHI, Hiroyuki NAKAGAWA, Tatsuhiro TSUC ...
2019 Volume 36 Issue 4 Pages
4_25-4_31
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
Software developers have made increasing use of natural language documents in many cases. Documents may contain useful information for software developers; however, it is difficult to extract such information when the number of the documents is considerably large. Latent Dirichlet Allocation (LDA) is a promising way of topic modeling. LDA-based topic modeling can be useful in facilitating comprehension of such documents. In LDA, a stop word list is used to filter general words for accurate topic classification. However, when using an existing stop word list, it is difficult to filter words that are not general but frequently appear in the target documents. In this paper, we propose a method that consists of two steps: stop word extraction from target documents and similar topic merging. We experimentally evaluate the method by applying it to mailing list. The experimental results demonstrate that our method constructs a topic model more accurately than the existing method.
View full abstract
-
Takuya KAWAKAMI, Hirohisa AMAN, Minoru KAWAHARA
2019 Volume 36 Issue 4 Pages
4_32-4_38
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
This paper proposes metrics for evaluating the risk of fault injection through a code change. The proposed metrics focus on the data dependence via variables and quantify the extent to which the code change would influence. The empirical study using 7 open source projects shows that the proposed metrics are useful explanatory variables together with conventional metrics in random forest models to predict fault injection commits.
View full abstract
-
Hiroaki KUWABARA, Yoshitoshi KUNIEDA
2019 Volume 36 Issue 4 Pages
4_39-4_45
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
This paper proposes bounded secrecy parameters in information flow analysis. Although secrecy parameters make it possible to define classes or functions without specifying a concrete secrecy for each data, programs that include secrecy parameters are required to satisfy noninterference with any substitution for secrecy parameters. Bounded secrecy parameters relax this too restrictive requirement and make more programs typable. We define a type system for information flow analysis of imperative programs with bounded secrecy parameters and show a simple example of type checking.
View full abstract
-
Kazuya TANAKA, Akito MONDEN, Yücel ZEYNEP
2019 Volume 36 Issue 4 Pages
4_46-4_52
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
auto-sklearn is a recent attention-gathering software library for automated machine learning that can automatically select appropriate prediction models and hyper parameters for a given data set. In this paper we empirically evaluate the effectiveness of auto-sklearn in software bug prediction. In the experiment, we used software metrics of 20 OSS projects for inter-version bug prediction and compared auto-sklearn with random forrest, decision tree and linear descriminat analysis by using AUC of ROC curve as a performance measure. As a result, auto-sklearn showed similar prediction performance as random forrest. We conclude that, although auto-sklearn is useful for bug prediction, we cannot expect better prediction performance than conventional modeling techniques.
View full abstract
-
Yu ADACHI, Haruto TANNO, Yu YOSHIMURA
2019 Volume 36 Issue 4 Pages
4_53-4_59
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
On a screen of the application, there are areas whose contents change every time the screen is displayed such as advertisements and news articles (hereafter, we call such areas dynamic content areas). When there are dynamic content areas, many differences in false positive are detected in Visual Regression Testing. Therefore, specifying the dynamic content area as the mask area and excluding it from the comparison area is effective for reducing the human labor to confirm these differences. We propose a method to specify a mask area based on the relative positional relationship of multiple screen elements and it can appropriately exclude dynamic content areas from the comparison area.
View full abstract
-
Kei ITO, Toshihiro KAMIYA, Akito MONDEN
2019 Volume 36 Issue 4 Pages
4_60-4_67
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Takeshi YOSHIMURA
2019 Volume 36 Issue 4 Pages
4_68-4_72
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Haruto TANNO, Futoshi IWAMA, Satoshi MASUDA
2019 Volume 36 Issue 4 Pages
4_73-4_78
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Michihiro HORIE, Koichi SASADA
2019 Volume 36 Issue 4 Pages
4_79-4_82
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
[in Japanese]
2019 Volume 36 Issue 4 Pages
4_84-4_85
Published: October 25, 2019
Released on J-STAGE: December 25, 2019
JOURNAL
FREE ACCESS
-
Koji TODA, Masateru TSUNODA
2019 Volume 36 Issue 4 Pages
4_95-4_106
Published: October 25, 2019
Released on J-STAGE: December 10, 2019
JOURNAL
FREE ACCESS
Machine learning has been commonly used to estimate the software development effort to assist project planning and/or management. Since project data sets for ML model construction often contain missing values, we need to build a complete data set that has no missing values either by using imputation methods. However, while there are several ways to build the complete data set, it is unclear which method is the most suitable for the project data set. In this paper, using project data of 1364 cases (34% missing value rate) collected from several companies, we applied six imputation methods (k-nn, applied CF, Miss Forest and Multiple Imputation method with three algorithm(Data Augmentation, Fully Conditional Specification and Expectation-maximization with Bootstrapping)) to build 6 machine learning methods. Then, using project data of 160 cases (having no missing values), we evaluated the estimation performance of each imputation method. Additionally, we evaluate an effect of log transformation to estimation accuracy. The result showed that k-nn was best performance in ideal situation, however if it can not be secured the estimation cost, multiple imputation with FCS algorithm is another choice.
View full abstract
-
Kazumasa SHIMARI, Takashi ISHIO, Katsuro INOUE
2019 Volume 36 Issue 4 Pages
4_107-4_113
Published: October 25, 2019
Released on J-STAGE: December 10, 2019
JOURNAL
FREE ACCESS
In this paper, we propose a method to collect an execution trace to analyze the internal states of software with a planned amount of storage. The method uses separated data buffers to record the actual values used by individual instructions. Changing the buffer size of each instruction, the method can record an execution trace using a limited size storage.
View full abstract