In Chapter 3, we introduced basic principles of cartography for mapping abstract structures commonly resulting from our thinking, ranging from concept mapping based on card sorting, through co-word maps derived from word co-occurrence analysis, to generic structures represented as networks, especially the interesting properties of a class of gigantic graphs known as small-world networks. We described typical dimensionality reduction techniques such as the classic multidimensional scaling (MDS) and the latest advances in non-linear MDS.

Chapter 4 [...] describes three interrelated aspects of science mapping: structural modeling, visual-semantic displays, and behavioraI semantics. StructuraI mapping is concerned with how to extract meaningful relationships from information resources. Visual-semantic displays focus on the design of effective channels for effective communication. Traditionally, structural mapping and visual-semantic display are regarded as the core of information visualization. Behavioral semantics emphasizes the meaning of behavioral pattems in helping us to understand the structure of an information space. It also provides a promising way to build responsive virtual environments. These enabling techniques are expected to play an increasingly important role in mapping scientific frontiers.