How to Adapt Models in a Dynamic World
This part of Data Science for Software Engineering: Sharing Data and Models explores ensemble learners and multiobjective optimizers as applied to software engineering. Novel incremental ensemble learners are explained along with one of the largest ensemble learning (in effort estimation) experiments yet attempted. It turns out that the specific goals of the learning has an effect on what is learned and, for this reason, this part also explores multigoal reasoning. We show that multigoal optimizers can significantly improve effort estimation results.
In summary, this chapter proposes the following data analysis pattern:
|Name:||Dynamic cross-company learning (DCL).|
|Also known as:||Online/dynamic ...|