So which classification method to use? There are generally two answers: (1) this (these) which work(s) best with your data and (2) as many as possible. The second makes the perfect sense because human perception works the same way, using all possible models until it reaches stability in recognition. Remember some optical illusions (e.g., the famous duck-rabbit image, Figure $$\PageIndex{1}$$) and Rorschach inkblot test. They illustrate how flexible is human cognition and how many models we really use to recognize objects.
At the end of the chapter, we decided to place the decision tree (Figure $$\PageIndex{2}$$) which allows to select some most important multivariate methods. Please note that if you decide to transform your data (for example, make a distance matrix from it), then you might access other methods: