Project without external funding

Application of Artificial Neural Network for the Prediction of Schering Stocks


Project Details
Project duration: 06/199612/2000


Abstract
The appplication of the classical statistical techniques for forecasting have reached their limit in the face of nonlinearities in the financial time series data. In the past years, it was shown that the Artificial Neural Networks are a superior tool in modelling these non-linear time series and therefore able to better describe the characteristics of stock markets.? ? The prediction of the movements in the stock market has currently emerged as an important research topic. Most research work there on ANN done in Germany based on the fundamentel analysis of stock market or on the technical analysis . Robles/Naylor [1996] showed the superiority of ANN compared with the traditional statistical models based on the technical analysis for copper trading. However they used for it only the weighted moving average . ? ? Siriopolos/Markellos and Sirlantzis [1996] used for the prediction of the short term trend of stock market general indeces for Gemany and Greece on ANN, a combination of - amongother items - the Price momentum and the Moving Average Convergence Divergence (which base on the structur of the Exponential Moving Average) as technical indicators. They found out, that due to the high volatility nature of stock market, the application of ANN is very useful.? ? ? The research done here is based on a time series of Schering stocks from January 1993 to February 1997. The aim is to develop a network configuratioon, is able to provide a good short term prognosis, what means the prognosis of the very next value. For a daily trader there is more benefit for having a reliable predictor in the short-term forecast - traditionally based in the technical analysis. Specifically the N2E4 neural network developed by the Research Group for Neural Networks at the University of Kassel on basis of technical indicators was used.


Principal Investigator

Last updated on 2017-11-07 at 14:11

Share link