|
[Previous] |
[Module] |
[Home] |
[List of Projects] |
[Next] |
| Project No. | 5003-039032 |
| Subject | Connectionistic information processing and financial market analysis |
| Author | Diethelm Würtz, PD Dr. Interdisziplinäres Projektzentrum für Supercomputing IPS, Swiss Federal Institute of Technology at Zurich (ETH)
|
| Start | 1.4.94 |
| End | 31.3.96 |
| Contents | It was long considered impossible to put financial market forecasting on a reliable footing. However, latest findings from the theory of nonlinear dynamic systems show that financial markets do conform to certain patterns of behaviour. Some interesting study results are available based for example on connectionistic models, and new analytical and forecasting systems are already operational. As a result, the use of new types of methods, above all from mathematics and engineering, could spell a radical change in financial-market analysis. The IPS researchers intend to draw up a feasibility study on the construction of a research prototype for the analysis and forecasting of financial markets. The aim is to develop a prototype showing the possibilities and limits of such a system.
The IPS has been involved in knowledge-based information processing, stochastic optimization methods, and in particular nonlinear methods of time sequence analysis, for many years. In 1990-92, its researchers laid the foundations for a system to analyse and predict financial market data on a working basis. The next step is to shorten the forecasting horizon and transition to tick data. Information-processing instruments for the analysis and forecasting of financial markets are of growing importance as decision-supporting tools. Potential beneficiaries of the IPS research are banks, insurance companies and consultancies. Contacts exist with firms in Switzerland and elsewhere. |
|
[Previous] |
[Module] |
[Home] |
[List of Projects] |
[Next] |