An international comparison of R and D activity, based on national statistics

The authors show what methods can be used to classify countries based on their statistics for research and development activity. The initial problem is that as many international tables can be compiled as there are indices employed. In the absence of standard statistical methods, selection among these becomes arbitrary. The solution recommended by the authors is to use compound tables compiled by analysis of main components and with genetic algorithms. They give examples of tables and point out the advantages and drawbacks of each method.