Background: The several maintenance tasks a system is submitted during its life usually cause its architecture deviates from the original conceivable design, ending up with scattered and tangled concerns across the software. The research area named concern mining attempts to identify such scattered and tangled concerns to support maintenance and reverse-engineering. Objectives: The aim of this paper is threefold: (i) identifying techniques employed in this research area, (ii) extending a taxonomy available on the literature and (iii) recommending an initial combination of some techniques. Results: We selected 62 papers by their mining technique. Among these papers, we identified 18 mining techniques for crosscutting concern. Based on these techniques, we have extended a taxonomy available in the literature, which can be used to position each new technique, and to compare it with the existing ones along relevant dimensions. As consequence, we present some combinations of these techniques taking into account high values of precision and recall that could improve the identification of both Persistence and Observer concerns. The combination that we recommend may serve as a roadmap to potential users of mining techniques for crosscutting concerns.