What are the disadvantages of SVM?

Can you elaborate on the drawbacks of using Support Vector Machines (SVM) in financial analysis and cryptocurrency predictions? Are there any limitations in terms of scalability, interpretability, or their ability to handle non-linear relationships effectively? Additionally, how might SVM's sensitivity to hyperparameters and the choice of kernel function affect the accuracy and stability of predictions in the volatile cryptocurrency market? Lastly, are there more suitable machine learning algorithms for this domain, and if so, why?