Analyzing a sustainability indicator by means of qualitative kernels

This chapter presents two totally new inference techniques for the field of interval and qualitative analysis, based on the kernel methods and statistical learning methodology, which have been shown to be very effective as machine learning. Moreover, on using them it is possible to work on any original space, and it is not necessary to work with numerical values (bioinformatics, string-type kernels). This study analyzes two approaches oriented at sustainability data: kernels on a discrete structure (the orders of magnitude) and interval kernels. Their good characteristics are studied with several examples, and an initial application to the sustainability data of the Town Council of Vilanova is performed. Finally, the feasibility is analyzed of a soft-computing tool which allows the degree of citizen satisfaction to be qualitatively assessed.