BACKGROUND: Several studies in software effort estimation have shown that it can be effective to use a window of recent projects, rather than the growing portfolio of all past projects, as training data. Practitioners need a method for choosing between those approaches when estimating effort for a target project. OBJECTIVE: This study examined the usefulness of the fitted values for choosing between moving windows and the growing portfolio. METHOD: An empirical experiment was conducted with a single-company ISBSG dataset. RESULTS: The fitted values could be useful for the selection on average. CONCLUSIONS: This positive result contributes to understanding when windows may be appropriate.
展开▼
