Partially supervised classification with optimal significance testing

The paper addresses the problem of estimating an optimal acceptance probability to be used for significance testing as applied to partially supervised classification where the class definition and corresponding training samples are provided a priori only for one specific class of interest. Considering the effort in both time and man-power required for a well-defined, exhaustive list of classes with their representative training samples even if there is just one class of interest to identify, the "partially" supervised capability would be very desirable, assuming adequate classifier performance can be obtained. The optimal acceptance probability is estimated directly from the data set. Experiments with both simulated and real data show very satisfactory results.<
>