I've done a fair bit of mathematical modelling over the years, and I think the comic market would be doable to an extent, but of doubtful utility. You'd need to have variables for the valuation of major market segments, and the rates of change of those could be pulled from data. You'd need to model couplings between market segments and estimate those ... blah blah blah.
Or you could just look it up here.
Edit: having skimmed the paper and reading key bits of it, they haven't approached it the way I would. But they are from a mathematics and statistics background and I was in the physical sciences. I'd be more interested in the market as a dynamic system of coupled variables, while their approach looks more like statistical regression at its heart, though with some dynamics through inputs such as 'deterioration rate' and even 'death rate of original purchasers' (!). In any case, the results are so-so (as I expected).