Anzeige
Erschienen in:
2017 | OriginalPaper | Buchkapitel
Statistical Models and Granular Soft RBF Neural Network for Malaysia KLCI Price Index Prediction
verfasst von : Dusan Marcek
Erschienen in: Advances in Time Series Analysis and Forecasting
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Two novel forecasting models are introduced to predict the data of Malaysia KLCI price index. One of them is based on Box-Jenkins methodology where the asymmetric models, i.e. EGARCH and PGARCH models were used to form the random component for ARIMA model. The other forecasting model is a soft RBF neural network with cloud Gaussian activation function in hidden layer neurons. The forecast accuracy of both models is compared by using statistical summary measures of model’s accuracy. The accuracy levels of the proposed soft neural network are better than the ARIMA/PGARCH model developed by most available statistical techniques. We found that asymmetric model with GED errors provide better predictions than with Student’s t or normal errors one. We also discuss a certain management aspect of proposed forecasting models by their use in management information systems.