This question already has an answer here:
What are some of the modern methods used to price equity volatilities "the most accurately possible" when there are very few listed derivative prices available or even none at all? Do the pricers in those cases resort to volatility forecasting based methods using historical data? Or do they try to use some similar equities that have listed prices and infer something from that?
I would appreciate a few references to the state of the art of the methods that are being used today and/or introductory material to this problem.