Beschreibung
Standard first-order Hidden Markov Models (HMMs) are very popular tools for the analysis of sequential data in applied sciences. HMMs are versatile and structurally simple models enabling probabilistic modeling based on a sound theoretical grounding. In contrast to the broad usage of first-order HMMs, applications of higher-order HMMs are very rare, but they have been proven to be powerful extensions of first-order HMMs including applications in speech recognition, image segmentation or computational biology. This book provides the first easily accessible and comprehensive extension of the algorithmic basics of first-order HMMs to higher-order HMMs coupled with practical applications in computational biology. The book starts with a theoretical part developing the algorithmic basics of higher-order HMMs and two novel model extensions (i) parsimonious higher-order HMMs and (ii) HMMs with scaled transition matrices. The second part considers applications of these models to the analysis of different DNA microarray data sets followed by a detailed discussion. The book addresses readers having basic knowledge on first-order HMMs interested to gain more insights on higher-order HMMs.
Autorenportrait
studied bioinformatics and received his doctoral degree from the Martin Luther University Halle-Wittenberg in 2010. He worked on plant computational biology and machine learning at the IPK Gatersleben and the IBENS Paris. Since May 2012, Michael Seifert is developing computational methods for cancer genomics at the BIOTEC TU Dresden.
