What is a suitable index to express similarity in two observations of the same set of variables containing ratios?

16 Views Asked by At

Is there an index that would express any kind of similarity between the values in the same variable sets of two observations, for example the distribution of energy-sources of two countries.

I work with a panel data of the energy mix of countries where each observation of a country has a set of variables for energy-sources expressed as shares of all energy consumed by origin. The variables are the same across panels and could for example be {coal, oil, nuclear, wind, hydro, other} with no missing data and {0.2, 0.2, 0.4, 0.1, 0.0, 0.1} as an example observation.

I wish to compare the similarity of the energy-mix of each country-pair expressed as a single measure, for example between 0 and 1.

Given whatever definition of similarity you find appropriate, how should I reason to compute a pair-wise similarity index in which two countries with a similar energy mix will score highly.