Regression in Estimation of Software Attributes: A Systematic Literature Review

Abstract

Software estimation is a fundamental activity in the Software development process, since it is possible to predict the number of defects, size, effort, among other attributes. With this, it is possible to improve the quality of the product and process. To predict quantitative values, it is common to use Regression Model mechanisms, although each model adjusts to a specific behavior of the data. In this work, a Systematic Literature Review is carried out based on the Kitchenham and Charters guide, to know the different types of Regression that have been used in the Software estimates. In addition, it seeks to know those attributes that are estimated and those that function as independent variables. Simple Linear Regression, Multiple Linear Regression and Logistic Regression were the most used, although other types of regression were found that can be further explored. Finally, the attributes that work as predictor variables were categorized, where the attributes of effort, lines of code and use cases were the most frequent.