The generation of profitable trading rules for Foreign Exchange (FX) investments is a difficult but popular problem. The use of Machine Learning in this problem allows us to obtain objective results by using information of the past market behavior. In this paper, we propose a Genetic Algorithm (GA) system to automatically generate trading rules based on Technical Indexes. Unlike related researches in the area, our work focuses on calculating the most appropriate trade timing, instead of predicting the trading prices.

In this paper, we propose a method to adapt sampling period of moving average dynamically and to analyze the kind of price changes. We also show ￣nding trends as experimental results of price °uctuation analysis using the proposed method.

This paper describes a prototype of financial trading strategy analysis system. It uses reinforcement learning to acquire a trading strategy and visualize the strategy. We illustrate how to use this system using an example for trading in the Japanese government bond market.

We present the use of Memetic Algorithms for the optimization of Financial Portfolios. Memetic Algorithms are hybrid algorithm where the evolution of individuals lead to the improvement of the portfolio structure, and local optimization rules contribute to the optimization of the weights of the financial assets. We compare this method with older GA-based methods for optimizing portfolios, and observe a noticeable improvement.

In this paper, we present an evaluation of a temporal rule generation method for trading dataset from the Japanese stock market. Temporal data mining is one of key issues to get useful knowledge from databases. To get more valuable rules for users from a temporal data mining process, we have developed a rule generation method which consists of temporal pattern extraction methods and rule induction algorithms. Using this method, we have done a case study to evaluate temporal rules from a Japanese stock market database for trading. Based on the result, we discuss about a way to utilize our rule generation method more effectively.

In this study, we proposed a new text-mining methods for long-term market analysis. Using our method, we analysed monthly price data of Japanese government bond market. First we extracted feature vectors from monthly reports of Bank of Japan. Then, trends of the JGB market were estimated by regression analysis using the feature vectors. As a result, determination coefficients were over 75%, and market trends were explained well by the information that was extracted from textual data. Finally, we compared the predictive power of textual data with that of numerical data. As a result, Our text mining method had prediction power superior to the numerical data analysis.

The purpose of this study is not predict a future stock price based on the past time series data, but is to clarify that how human predict a future stock price, when they see the past time series da Our experiments use computer program that shows time series data on discrete graph to subjects. It is considered that these time series data ( 1 ? 15 day or 1 ? 30 day ) are the past stock price. Subjects are required to predict future price of this stock at certain future time (31st, 35th , 45th , 55th day). The result of this experiment is that human adopt two ways of prediction. The first way is strong depend on nearest past data, man use this way to predict very near future. The second way is almost linear regression using all given data. However this is not just linear regression, but their prediction is approached to average of the past data.

I constructed the stock artificial market in order to analyze why an activist's announcement of the information that he holds over 5% of the firm's stock brings about rising of the stock price. I prepared two hypothesis. The first hypothesis is that the stock price rises because of speculators who consider the announcement to be the sign that the firm is undervalued. the second hypothesis is that investors expect jawboning rather than proxy contest. As a result, I found that the expectation for jawboning is much larger than proxy contest. Many investors regard the effect of jawboning as important.

We simulate stock returns using the artificial market using agents managing portfolio on real market data. In this model to couple real markets and an artificial market, cash inflows (outflows) to (from) agents are estimated by stock returns of the real markets. Our model can predict stock returns on some level.

In financial trading, total order amount changes by Trader's order. And, it changes various speed. In this article, we'll be showed equilibrium formula in this dynamics by classical Operations Research model. Originally, this model was used for military purpose. Now, we'll use this model to find dynamic equilibrium on Buyer's and Seller's "Market Share". And, we'll get effect by execution precision at this dynamic equilibrium.