Proceedings Fourth International Conference on Computational Intelligence and Multimedia Applications. ICCIMA 2001, 2001
We discuss the stylized facts which are observed in the log return of almost all financial market... more We discuss the stylized facts which are observed in the log return of almost all financial markets. We especially consider an interacting agent model of collective opinion formation in a stock market. For this purpose. a two-dimensional Ising model has been applied. The evaluation of the model suggested some key points to improve simulation precision
We conducted an empirical investigation regarding the effects on the eigenvalue elicited from the... more We conducted an empirical investigation regarding the effects on the eigenvalue elicited from the random matrix theory (RMT) as the number of stocks that consist a correlation matrix increases. An additional objective is to examine whether the properties of eigenvalue change according to the number and types of stocks in the correlation matrix. According to the observed results, we found that the magnitude of the eigenvalue elicited via the RMT method clearly increases in proportion with the number of stocks in the correlation matrix. Furthermore, we discovered that the largest eigenvalue maintains its identical properties regardless of the changes in the number and types of stocks in the correlation matrix, whereas other eigenvalues deviated from the range of a random matrix display different properties.
ABSTRACT Symbolic analysis of time series of economic indicators offers an advantage of transferr... more ABSTRACT Symbolic analysis of time series of economic indicators offers an advantage of transferring quantitative values into qualitative concepts by indexing a subset of intervals with a set of symbols. In a similar way, computer codes routinely process continuous problems in a discrete manner. This work explains an appealing analogy between the DNA code of life and the symbol series derived from financial markets. In particular, it is shown that similarity scoring schemes and the alignment gap concept known in bioinformatics have even more natural and deeper analogies in the economic systems. The symbolic analysis does not solely mean a loss of information; in also allows us to quantify a similarity degree between various financial time series (and their subsequences) in a rigorous way, which is a novel concept of practical importance in economic applications. Our symbolic analysis concept is illustrated by two types of market indicator series, namely the analysis of Dow Jones vs. NIKKEI 225 indices on one side, and the CZK/EUR exchange rate vs. Prague money market rates on the other side. The present framework may also yield a significantly reduced computational complexity as compared to the neural networks in the class of similarity-comparison algorithms.
So far econophysics has given contributions in four areas of economics: financial markets, wealth... more So far econophysics has given contributions in four areas of economics: financial markets, wealth and income distribution, industrial economics (firms’ size distribution, growth rates) and, more recently, networks analysis. According to Gallegati et al., 2006, there are some weakness in the approach: a lack of awareness of work that has been done within economics itself; resistance to more rigorous and
Proceedings Fourth International Conference on Computational Intelligence and Multimedia Applications. ICCIMA 2001, 2001
We discuss the stylized facts which are observed in the log return of almost all financial market... more We discuss the stylized facts which are observed in the log return of almost all financial markets. We especially consider an interacting agent model of collective opinion formation in a stock market. For this purpose. a two-dimensional Ising model has been applied. The evaluation of the model suggested some key points to improve simulation precision
We conducted an empirical investigation regarding the effects on the eigenvalue elicited from the... more We conducted an empirical investigation regarding the effects on the eigenvalue elicited from the random matrix theory (RMT) as the number of stocks that consist a correlation matrix increases. An additional objective is to examine whether the properties of eigenvalue change according to the number and types of stocks in the correlation matrix. According to the observed results, we found that the magnitude of the eigenvalue elicited via the RMT method clearly increases in proportion with the number of stocks in the correlation matrix. Furthermore, we discovered that the largest eigenvalue maintains its identical properties regardless of the changes in the number and types of stocks in the correlation matrix, whereas other eigenvalues deviated from the range of a random matrix display different properties.
ABSTRACT Symbolic analysis of time series of economic indicators offers an advantage of transferr... more ABSTRACT Symbolic analysis of time series of economic indicators offers an advantage of transferring quantitative values into qualitative concepts by indexing a subset of intervals with a set of symbols. In a similar way, computer codes routinely process continuous problems in a discrete manner. This work explains an appealing analogy between the DNA code of life and the symbol series derived from financial markets. In particular, it is shown that similarity scoring schemes and the alignment gap concept known in bioinformatics have even more natural and deeper analogies in the economic systems. The symbolic analysis does not solely mean a loss of information; in also allows us to quantify a similarity degree between various financial time series (and their subsequences) in a rigorous way, which is a novel concept of practical importance in economic applications. Our symbolic analysis concept is illustrated by two types of market indicator series, namely the analysis of Dow Jones vs. NIKKEI 225 indices on one side, and the CZK/EUR exchange rate vs. Prague money market rates on the other side. The present framework may also yield a significantly reduced computational complexity as compared to the neural networks in the class of similarity-comparison algorithms.
So far econophysics has given contributions in four areas of economics: financial markets, wealth... more So far econophysics has given contributions in four areas of economics: financial markets, wealth and income distribution, industrial economics (firms’ size distribution, growth rates) and, more recently, networks analysis. According to Gallegati et al., 2006, there are some weakness in the approach: a lack of awareness of work that has been done within economics itself; resistance to more rigorous and
Uploads