A Note on Modeling of Quality Evaluation Based on Large Data Sets in Software Development Projects  

Oleh:

Morita, Takayuki ; Esaki, Kazuhiro ; Kimura, Mistuhiro

Jenis: Article from Proceeding
Dalam koleksi: The 14th Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS), 3-6 December 2013 Cebu, Philippines, page 1-8.
Topik: Software Quality; Project Management; Multiple Regression Model; Real Data Analysis
Fulltext: 1216.pdf (240.74KB)

Isi artikel

At present, it is widely-recognized that the quality of large-scale software system is not better than that of hardware. Thus software development processes still have a lot of issues to be solved. Many researchers and practitioners have often discussed how to improve software quality. Lots of quality evaluation methods on the software development process have been proposed based on try and error schemes. In this study, we produce software quality evaluation models by using the multiple regression analysis from several software development projects. IPA/SEC (Information-technology Promotion Agency/ Software Reliability Enhancement Center) has been collecting these data sets from about 30 companies since 2004, and the number of collected projects is over 3000. Since the collected raw data sets contain some inappropriate (e.g. incorrect missing) data, we firstly cleaned up data sets. On the multiple regression analysis, we defined that the objective variable was the number of software bugs per program volume measured by kilo lines of code, and the candidates of explanatory variables consisted of 611 attributes of software development processes. To reduce the number of explanatory variables, we additionally define 34 attributes. These variables consist of quantitative and qualitative attributes. In the actual multiple regression analysis, we chose several explanatory variables from the 34 attributes by the stepwise method. Consequently, the number of data sets is reduced to 94 in the best model before performing the stepwise method. As a result of the regression analysis, we found that the quality of developed software is subject to influence of the four factors. In conclusion, we summarize that managing these attributes in the planning stage of software development would improve the quality of software.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali