If I ever read a book with a misleading title, that would be it.
Not because there is no Clojure in it, but the amount of Clojure used is minimal compared to the whole.
Also, not because there is no Machine Learning in it, but the book goes as deep as explaining the algorithms behind each common machine learning technique, without explaining when you should use it. There is ample discussion about the mathematical context of each method, but it explains absolutely nothing about machine learning itself -- it's purely a bunch of mathematical equations that could be use to extract some pattern, but it's hardly "machine learning" at this point.
(Also: Neural networks as "unsupervised learning"?!?)
So, good book on the math behind some machine learning equations, very bad clojure, very bad machine learning.