Maximum-likelihood approach for signal estimation in direct detection experiments for Dark Matter

We develop a maximum-likelihood method for estimating the signal in direct Dark Matter detection experiments which rely on one or more parameters for discriminating against background. We demonstrate that the method performs better than the traditional “cut” approach, at least in the limit of large number of events. We apply the method to the CDMS II data acquired at the shallow underground site, and we show that the method could improve the sensitivity of the experiment by 15-20%, as compared to the sensitivity relying on the cut in the discriminating parameter.