An alternative approach for building composite indicators using entropy and multicriteria decision analysis
Orador: Doutor António Xavier
Abstract:
Composite indicators and the resulting rankings are a useful tool since they allow for simplifying complex reality and easy comparisons among territorial units. To build composite indexes, Multicriteria Decision Analysis (MCDA) provides a considerable set of methods that may be used in different areas of research. Regarding composite indicators and ranking analyses, two relevant less used approaches are: Extended Goal Programming and Compromise Programming. These approaches have several methodological advantages, nevertheless, they encountered some problems: 1) The construction of the ranking, since it is an iterative process, took too much time for a large number of units; 2) Several situations were treated as having the same ranking position, despite being, in reality, different. To overcome these problems, the proposed methodological approach introduced the concept of normalized entropy in these methods. The remainder of this seminar is presented as follows. Firstly, the main critical factors regarding composite indexes construction are analysed. Then, the Extended Goal Programming and Compromise Programming approaches, with the entropy improvements are presented. Finally, several case studies, are discussed.