Figure 31. Non-linear map (stress 0.8%) derived from the distance matrix in Table 6, showing the approximate relationships between the muscle biopsy spectra of dystropic patients (D,E,F) and controls (A,B,C) in 40-dimensional space (see text).
the samples. Furthermore, non-linear mapping can provide a unique insight into the influence of varying the analytical conditions, e.g. pyrolysis temperatures, on the degree of differentiation achieved. Finally, it should be noted that Kruskal's mapping technique does not require any assumptions about the statistical distribution of the data. This allows mapping of small data sets with insufficient data for a reliable principal component, factor analysis or discriminant analysis approach.
Was this article helpful?