An Ensembel Model for Modelling Chaotic Behaviour of Bursa Malaysia Time Series Data

Financial data is characterized as non-linearity, chaotic in
nature and volatility thus making the process of forecasting cumber-
some, hence a successful forecasting model must be able to capture long-
term dependencies from chaotic data. In this study, an ensemble model,
called UKF-NARX, consists of unscented kalman �lter and parallel non-
linear autoregressive network with exogenous input trained with bayesian
regulation algorithm is modelled for chaotic �nancial forecasting. The
proposed ensemble model is compared with the conventional non-linear
autoregressive network and �nancial static forecasting model employed
by �nancial analysts when applying in multi-step-ahead forecasting. Ex-
perimental results on Burssa Malaysia KLCI show that the proposed
ensemble model outperforms the other two commonly used models.