Stock Market Modeling and Forecasting translates experience in system
adaptation gained in an engineering context to the modeling of financial
markets with a view to improving the capture and understanding of market
dynamics. The modeling process is considered as identifying a dynamic
system in which a real stock market is treated as an unknown plant and
the identification model proposed is tuned by feedback of the matching
error. Like a physical system, a financial market exhibits fast and slow
dynamics corresponding to external (such as company value and
profitability) and internal forces (such as investor sentiment and
commodity prices) respectively. The framework presented here, consisting
of an internal model and an adaptive filter, is successful at
considering both fast and slow market dynamics. A double selection
method is efficacious in identifying input factors influential in market
movements, revealing them to be both frequency- and market-dependent.
The authors present work on both developed and developing markets in the
shape of the US, Hong Kong, Chinese and Singaporean stock markets.
Results from all these sources demonstrate the efficiency of the model
framework in identifying significant influences and the quality of its
predictive ability; promising results are also obtained by applying the
model framework to the forecasting of major market-turning periods.
Having shown that system-theoretic ideas can form the core of a novel
and effective basis for stock market analysis, the book is completed by
an indication of possible and likely future expansions of the research
in this area.
