APPLICATION OF PATTERN RECOGNITION TO SOME PROBLEMS IN ECONOMICS

Abstract

The choice of a collective decision rule by some outside "advisor" on the basis of various rationality requirements is the classical problem of "individual preference aggregation" in economics. The logic aim of this study is to show how one can use a single underlying structure, viz. that of a preference pattern, and its various mathematical representations to tackle the problem of social choice. The use of classical methods of pattern recognition was used and it led to a completely general concept of collective decision rule and actual processes for collective decision making. We reviewed the aggregation problem for consumer preferences and then translated its formulation in terms of pattern recognition which gave us a very clear understanding of the so-called "voting paradox." Then we related some classical methods of pattern recognition with the notion of a collective decision rule. And finally, we present two possible aggregation algorithms for binary preference patterns. The results thus obtained cast even more light on the relation between pattern recognition and the aggregation problem.