Sarkar, Atri2016-04-272016-04-272016-04-272016-04-22http://hdl.handle.net/10012/10406A key challenge of the development and maintenance of configurable systems is to predict the performance of individual system variants based on the features selected. It is usually infeasible to measure the performance of all possible variants, due to feature combinatorics. Previous approaches predict performance based on small samples of measured variants, but it is still open how to dynamically determine an ideal sample that balances prediction accuracy and measurement effort. In this work, we adapt two widely-used sampling strategies for performance prediction to the domain of configurable systems and evaluate them in terms of sampling cost, which considers prediction accuracy and measurement effort simultaneously. To generate an initial sample, we develop two sampling algorithms. One based on a traditional method of t-way feature coverage, and another based on a new heuristic of feature-frequencies. Using empirical data from six real-world systems, we evaluate the two sampling algorithms and discuss trade-offs. Furthermore, we conduct extensive sensitivity analysis of the cost model metric we use for evaluation, and analyze stability of learning behavior of the subject systems.enperformance predictionconfigurable systemssampling techniquesMaster Thesis