ABSTRACT
In some applications, when a large number of types of objects in a relational data set are related to each other in a complicated way, we may want to focus on a certain type of data objects to reduce the model complexity. In this situation, the same type of data objects is described by different sets of features and different sets of relations. Hence, the relational data clustering problem can be viewed as a case of different view learning in this situation. In this chapter, we propose a general model for multiple-view unsupervised learning. The proposed model introduces the concept of mapping function to make the different patterns from different pattern spaces comparable and hence an optimal pattern can be learned from the multiple patterns of multiple representations.
