Billio, M.

model density forecast weight combination scheme prediction yt +1 forecasting survey time t density forecasts strategy value journal combination scheme return method time-varying distribution fferent dynamic stock result probability parameter time-varying weights point combination schemes section bayesian application predictor approach example point forecasts paper portfolio error vector speci series figure process 0.2 sequential monte carlo rmspe square prediction error survey forecasts multivariate period uncertainty carlo accuracy monte 0.5 utility flation 0.4 nonlinear bayesian model weight dynamics variable score space panel order university performance sample sequential di fferent models ∝ exp − 0.6 index state investor analysis fi nancial crisis noise model