Discriminant Analysis in Bilinear ProcessesUsing State Space Methodology

This pap er is concerned with discrim inating betweentwo bilinear t ime series processes. the classical methods to discrimination makes stationary and linearity assumptions to pr ocesses. however , in general, bilinear pro cesses are not stationarynor linear . the method suggeste d in t his pap er is based onstate space methodology and uses cumulat ive sum cont rol cha rt .classificat ion of a ser ies to hi or h 2 is t reated like a sequentiallikelihood ratio test where optimal decision is to be made as soonas possible.the method has been carr ied out to bilinear processes, but can beused to any adopted pro cesses to state space models.