Hello all,I am trying to model the mean reversion parameter lambda from the Ohrenstein-Uhlenbeck SDE ( dS = lambda ( S - Sbar ) dt + sigma dW. I have a large time-series to work with. Any guidance is appreciated. Thanks.

I would probably start by searching these forums for "mean reversion calibration" as I recall seeing many threads dealing with this over the past few years. Dixit and Pindyck also discuss methods in their "Investment Under Uncertainty". I also recall seeing many articles related ot calibrating mean reversion parameters written by Clewlow and Strickland, primarily for the commodity sector. Try googling "Clewlow Strickland Mean Reversion" and I think you will have some luck. Generally, regression is a good start, but other methods like MLE can yield better estimates.