An empirical study to develop a Decision Support System (DSS) for measuring the impact of quality measurements over agile Software Development (ASD)

Abstract

Background/Objectives: Primarily, this quantitative research aims to study the impact of integrating quality measurements with ASD, quantify it, and develop a DSS for predicting its outcome. Methods/Statistical Analysis: Included within a survey, the population sample is represented by project managers, who were divided into two independent groups: The first one adopts an explicit quality measurement framework while the second group does not apply quality measurements. After that, the researcher tested both groups in an independent samples t-test, and analysed results statistically. After experimenting different machine learning models, the researcher developed a DSS based on Linear Regression. Findings: Only 150 responded out of 200 respondents. The research dataset passed the "independent t-test" validity test with the fulfilment of the six assumptions. After conducting the independent t-test design, the researcher found that the value of Sig. (2-tailed) is less than .05, which means that the differences between the experimented groups are statistically significant. After that, the researcher utilized WEKA experimenter with 10-folds cross validation to test the dataset fitness with four different machine learning algorithms, which are Linear Regression (base), Multilayer Perceptron, KStar, and Decision Stump. The results showed that Linear Regression (base) provides better fitness with the dataset. Moreover, The R Square for it is .836. Based on Linear Regression, the researcher developed web and windows version of the DSS using VB.NET. In summary, research results shows that there is empirical evidence to support the proposition that quality measurements integration with ASD presents a strategic value to organizations. The contribution of these findings is materialized in its empirical nature and the scariness of research in this domain. Application/Improvements: Henceforward, the researcher are planning to expand the population sample, publishing the developed DSS online with integrated feedback, and developing other DSSs for supporting integrating quality measurements with ASD.