I have data on the metallic compositions (percentages of specific elements) of 30 coins, each minted in a different year from 1902 to 1967. I also have the composition of a coin minted in an unknown year which I would like to try and classify to a year of minting through comparison with the reference dataset.

How would I go about doing this?

I attempted discriminant analysis in SPSS using Coin Year (numeric; nominal) as the grouping variable and the percentages of each metal element (Cu, Sn, Zn, Pb, As, Zr, Ni, Nb, Co, W, Mn, Se) as the Independents, but received the error: “Warnings: There are 31 non-empty groups and 31.000 (31 unweighted) cases that are valid. Not enough weighted or unweighted cases. Execution of this command stops.” I suspect this is because there is only one set of values associated with each year, but I’m stuck on how to overcome this.

I thought about potentially using a Euclidean distance measure, but I’m not sure. Any advice?