GUI test case migration is to migrate GUI test cases from a source app to a target app. The key of test case migration is widget matching. Recently, researchers formulate widget matching as a matching task. However, since these matching approaches depend on static word embeddings without utilizing contextual information to represent widgets and manually formulate matching functions, the approaches may encounter difficulties when handling complex matching relations in real-world apps.

In this paper, we propose the first learning-based widget matching approach named \emph{TEMdroid} (\underline{\emph{TE}}st \underline{\emph{M}}igration) for test case migration. Unlike existing approaches, TEMdroid uses BERT to capture contextual information and learns a matching model to match widgets. Additionally, to balance the significant imbalance between positive and negative samples in real-world apps, we design a two-stage training strategy where we first train a hard-negative sample miner to mine hard-negative samples, and further train a matching model using positive samples and mined hard-negative samples. Our evaluation on 34 apps shows that TEMdroid is effective in event matching and test case migration. For event matching, TEMdroid’s Top1 accuracy is 76%, improving over 17% compared to baselines. For test case migration, TEMdroid’s F1 score is 89%, also a 7% improvement compared to the baseline approach.