I'm exploring the limitations of Support Vector Machines (SVM). Could you elaborate on the common issues or problems associated with SVM, such as sensitivity to parameter settings, difficulty in interpreting results, or any other relevant challenges?