As observed in Chapter 5, few methods have been proposed to fuse the components of an ensemble of topology-preserving maps into a final map that summarizes the information of the different composing maps(Georgakis et al, 2005; Saavedra et al, 2007). Nevertheless, as previously explained, there are multiple advantages to this approach.
Seeking further improvements to the results, two different approaches have been followed and are presented in the rest of this chapter. Firstly, an algorithm combining two of the previously proposed models -Fusion by Distance and Fusion by Similarity- is presented in Section 6.2.
Subsequently, an improved version of the Superposition algorithm is presented in Section 6.3.1.