RQ1: To what extent can our approach reduce the number of invalid refactorings compared to other multi-objective refactoring recommendation techniques?