This preview shows page 1. Sign up to view the full content.
Unformatted text preview: the future stock prices. Since, 2 this is also the focus of the present project, some of these papers are described in greater
detail especially with respect to the feature extraction, wherever such information has been
made available in these publications.
Modular neural network approach
Kimoto et al  use several neural networks trained to learn the relationships between
past values of various technical and economic indices for obtaining the expected returns of
the TOPIX. The TOPIX is a weighted average of all stocks listed on the Tokyo Stock
Exchange and is similar to the Dow Jones Industrial Average (DJIA). The technical and
economic indices used are: the vector curve (an indicator of market momentum), turnover,
interest rate, foreign exchange rate and the DJIA value. The desired output of the
networks is a weighted sum, over a few weeks, of the logarithm of the ratio of the TOPIX
at the end of week t, to the TOPIX value at the end of week (t-1).
Thus, rt = ln(TOPIX(t)/TOPIX(t-1)) and the desired output is a weighted sum...
View Full Document
- Spring '12