There is a lot of material that needed to be covered in this text. The ﬁrst part on fundamential material was necessary in order to properly understand the Bayesian Source Separation model. I have tried to provide a coherent description of the Bayesian Source Separation model and how it can be understood by starting with the Regression model. Throughout the text, Normal likelihoods with Conjugate and generalized
Conjugate prior distributions have been used. The coherent Bayesian Source Separation model presented in this text provides the foundation for generalizations to other distributions. As stated in the Preface, the Bayesian Source Separation model incorporates
available prior knowledge regarding parameter values and incorporates it into the inferences. This incorporation of knowledge avoids model and likelihood constraints which are necessary without it.