Extreme value theory and neural network for catastrophic fall prediction: a study of year 2008-2009 Online publication date: Sat, 05-Jul-2014
by Utkarsh Shrivastava; Gyan Prakash; Joydip Dhar
International Journal of Information and Decision Sciences (IJIDS), Vol. 6, No. 2, 2014
Abstract: Extreme value theory to a certain extent is successful in modelling extreme events as it assumes that outliers follow distribution other than normal. However, a mathematical model might not just be sufficient to predict extreme events. Nowadays, extreme events have become so common that investors' past experience of such situations takes over and plays important role in guiding collective behaviour during downturn. In this study, firstly extreme events are modelled using generalised extreme value (GEV) distribution. Secondly, past deviations from return levels obtained as quantile of GEV distribution and future risk of market falling below the same level are classified using perceptron network. Neural network is basically used to inculcate learning form past market movements to predict future. Trained network hence obtained is used for simulating monthly risk for catastrophic years 2008 and 2009. Comparison of actual and forecasted results indicates substantial improvement in market fall prediction.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Information and Decision Sciences (IJIDS):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com