The Effects of Gradual Weighting on Duration-Based Moving Windows for Software Effort Estimation

Several studies in software effort estimation have found that it can be effective to use a window of recent projects as training data for building an effort estimation model. Windows can be defined as having a fixed size (containing a fixed number of projects), or as having a fixed duration. A recent study extended the idea of windows, by weighting projects differently according to their order within the window, and found that weighted moving windows could significantly improve estimation accuracy. That study used fixed-size windows. This study examines the effect on effort estimation accuracy of weighted moving windows that are based on fixed duration. We compare weighted and unweighted moving windows under the same experimental settings. Weighting methods are found to improve estimation accuracy significantly in larger windows, and the methods also significantly improved accuracy in smaller windows in terms of MRE. This result contributes further to understanding properties of moving windows.