Improving actionable warning identification via the refined warning-inducing context representation

We improve AWI via the refined warning-inducing context representation, which captures both lexical and structural information for AWI from the refined warning-inducing context. We conduct experiments on over 51K+ warnings from 56 releases of five large-scale and open-source projects. The results in both within-project and cross-project AWI show that our approach is more effective than four state-of-the-art ML-based AWI approaches.