Work has been done on using Markov models etc to predict missing symbols in these texts. But it feels like with all the data now available, and the fact that so

Work has been done on using Markov models etc to predict missing symbols in these texts. But it feels like with all the data now available, and the fact that some signs' meanings are known, we must be able to at least reduce the constellation of possible meanings of some of the unknown signs. There are only so many things it was possible to say about olives in the ancient world, and presumably the semantic space wouldn't be so different to other vaguely contemporaneous languages (not just Linear B). Does anybody know of any work in this direction?