Hyvarinen and fellow researchers Juhu Karhunen and Erkki Oja (all Helsinki U. of Technology) introduce independent component analysis as a statistical and computational technique for revealing hidden factors that underlie sets of random variables, measurements, or signals. Readers are intended to be from such disciplines as statistics, signal processing, neural networks, information theory, and engineering, and to have a grounding in college calculus, matrix algebra, probability theory, and statistics. Exercise problems and computer assignments facilitate the book's use in a graduate course.

Читать «Independent Component Analysis»

Купить электронную версию