We investigate whether the number and content of postings of the posting of BBS become the explanatory variables of the stocks return. In previous work, the possibility that these factors are related to the stocks return and risk is suggested. We verify whether CAPM approve a portfolio by the number and contents of postings in TSE. We find the possibility CAPM is not approved in a portfolio with the highest number of postings and the fewest number of postings as well as the highest number of bullish postings and the highest number of bear postings. We argue an index of the difference of the return of a portfolio with the highest number of postings and the return of a portfolio with the fewest number of postings as well as the posting contents.
In this study, we proposed a new text-mining methods for long-term market analysis. Using our method, we perfomed out-of-sample test using monthly price data of financial markets; Japanese government bond market, Japanese stock market, and the yen-dollar market. First we extracted feature vectors from monthly reports of Bank of Japan. Then, trends of each market were estimated by regression analysis using the feature vectors. As a result, As a result, the method could estimate JGB market best and the stock market is the second.
Since the subprime mortgage crisis in the United Sates, stock markets around the world have crashed, revealing their instability. To stem the decline in stock prices, short-selling regulations have been implemented in many markets. However, their effectiveness remains unclear. In this paper, we discuss the effectiveness of short-selling regulation using artificial markets. An artificial market is an agent-based model of financial markets. We constructed an artificial market that allows short-selling and an artificial market with short-selling regulation and have observed the stock prices in both of these markets. We found that the market in which short-selling was allowed was more stable than the market with short-selling regulation, and a bubble emerged in the regulated market. We evaluated the values of assets of agents who used three trading strategies, specifically, these agents were fundamentalists, chartists, and noise traders. The fundamentalists had the best performance among the three types of agents.
This paper constructs the option market model and analyzes the real market with evolutionary game theory. This model can predict the next market states with the equilibrium stability condition. In addition, this paper compares this model and Black and Sholes [1]. We can interpret that this model gives a player's micro-foundation with Black and Sholes [1].
Various market studies have made such as micro market structure. I want to pick up a case study applying optimization of decision making with extracting professional traders decision making process through survey. To create the ideal model of decision making by AI are extremely interesting. As a background of this study, behavioral finance theory such as Prospect Theory(Daniel Kahneman and Amos Tversky 1979)are used.
It is reported that, in individual stocks, a large increase in trading volume indicates stronger return persistence or weaker reversal effects. A reason given for this volume-return relation has been that it can signal the existence of important fundamental news. In this study, we present another plausible explanation. Through empirical analysis, we show that the volume-return relation remains strong even if there is no important fundamental news. Applying market model simulation, we demonstrate that investors' trend-chasing behavior can cause this volume-return relation. Our findings suggest that the relation can be caused by factors not directly related to fundamental news, especially by investors' trend-chasing behavior.
Recently the artifcial markets are attracting attention of many researchers. To improve the reliability of the artifcial market, it is required to show that the artifcial market reproduce the real market. In this paper, we propose a method to calculate similarity between behaviors of the markets. In this method, we focus on the board data instead of the price data, and perform time-series analysis. Our simulation shows the efectiveness of the proposed method. In addition, we confrm the market difference between before and after the global depression using our method.
This study proposes methods to quantify states of market participants with comprehensive high-resolution data of financial markets. The number of quotations and transactions are focused from information transmission perspectives. The proposed methods are demonstrated by using comprehensive high-resolution data of the foreign exchange market. It is found that the states of market participants drastically changed around recent financial crisis which had happened in the middle of 2008.
Porfolio Optimization is an important problem for financial engineering. It consists of finding out the best investment weights for a large group of assets, so that the Expected Return of those assets is maximized and the specific risk of the portfolio is minimized. This problem can be modeled as a Parameter Optimization problem, and Genetic Algorithms have shown better results every year. In this paper we review recently proposed techniques to optimize Financial Portfolio using Historical Price values, compare them, and draw up proposals about how to improve these results even further.
This paper describes a framework of reinforcement learning for finance. I propose a new reinforcement learning algorithm based on Q-learning. I show the experimental results using N-arms bandit and gridworld problems.
In this paper, we proposed automated trading system using Genetic Algorithm(GA) and Genetic Programming(GP). There are many reports which focus on system trading. In recent years, these researches have attracted attention because of impact of finance crisis. Among them, we focus on two area, one is optimization of technical indicators, the other is optimization of technical indicators combination. There is not research which focuses on both parts. So we proposed automated trading system by optimizing indicators and combination of them. At first, their parameters are optimized by GA. Each indicator is fitted on movement of current market by this process. Then, using optimized indicators, combination of them is optimized by GP to generate buy-sell strategy. To verify effectiveness of proposal method, we simulated using real data given by Gaitame.com. Consequently, we got good result and verified effectiveness of proposal method.