本論文では、オンライン・ポートフォリオアルゴリズムが現実的にどの程度有効であるかを検証した。我々は、理想的な環境や選択株式を主体として極めて高い効用を示している手法を、より現実的なデータセットと、実践的評価による検証を行った。これらの手法が、長期に安定した運用が行えることを最重要とし、単なる効用の最大化ではなく、現実に起こりえる問題を想定した複数の評価指標が不可欠と考えた。6つの評価指標を提案し、効用を実践的な優位性として評価できる仕組みを構築する。評価手法は代表的なアルゴリズムであるAntiCor および最新の手法であるOLMAR である。これらの手法を、2000~2014 年の期間における米国株のNasdaq100 から69 銘柄・SP500 から413 銘柄で検証したので報告する。
We introduce two kinds of arbitrage opportunities in the foreign exchange market by using high-frequency data over 12 years; one is negative spread arbitrages and the other is triangle arbitrages. We already showed the relationship between the occurrence of the arbitrage and volatility, the number of deals and the number of computer traders[1]. In this paper, we take into account an execution risk in our arbitrage analysis. We found the expected profit of the arbitrage declines year by year. We assume this result caused by increasing of the algorithm traders which can detect arbitrage opportunities much faster than human traders.
Financial crises are typically caused by a chain of credit contractions, which in turns could be caused by the rapid worsening of indexes that indicate people's psychology, such as bank stock prices. The purpose of this analysis is to identify trigger points where bank stocks rise or fall by extracting what common points existed in financial economic indicators immediately before significant fluctuations of bank stocks occurred in the past. To conduct discriminant analysis, we used the traditional statistical method as well as ensemble learning. We also used "bank stock performance" as well as "bank stock regime change" as objective variables. This attempt showed that the money multiplier and 10-year yield of government bonds are important ones that could have an influence on bank stock regime change. Keywords : Ensemble Learning, Bank Stock, Regime, J48, Random Forest.
サブプライム問題やリーマンショックのような金融危機では,ある金融市場の混乱が他の市場に伝播していくことがある.その理由の1つとして,ある市場で生じたショックによりポートフォリオのリバランスが発生し,その結果他の金融市場の価格にも影響を与えることが挙げられる.これまでにショック伝播発生メカニズムを解明する研究は行われてきたが,金融ショックの伝播が発生しない金融市場に関してはほとんど着目されていない.そこで,本研究では,人工市場を用いて資産下落の伝播が起こらない市場要因は存在するのか調査した.その結果,資産の信用売りが,信用買いや現物買いと同じくらい活発に行われる市場では資産下落の伝播は起こらない可能性があることを発見した.
There are many researches to investigate mechanism of financial markets using artificial market model. Yamada et al. construct the stochastic dealer model for financial markets. This model achieved remarkable results. But they don't take volume of orders into account. Then we construct a stochastic dealer model taking the volume of orders into account.
We built an artificial market model and investigated for market maker's impact to the competition among markets. If there is a market maker in a minor market, we found that share of the major market's volume transfered to the minor market. Although market maker's spread was wider than the average of major market's bid-offer-spread, share of the major market's volume transfered to the minor market and the speed of transfer from the major market to the minor market depended on the market maker's spread. We also analyzed the mechanism and revealed it.
Investing in single or similar financial instruments is dangerous from the viewpoint of risk. In order to realize wide range of investment , make the clustering focuses on stocks companies that Investment Trusts are investing in this study. And I examine a wide range of investment can help to diversification.
Recent research has explored the proper method to analyse the relationships in financial markets for risk management. In this paper, we applied and expanded the method of transfer entropy to construct the influence network which represents the information propagation between stocks. FIrst, we demonstrate that this method reveals meaningful hidden relations of cause and effect between stocks, by showing that it is more useful than other methods, such as normal Transfer Entropy network, correlation coefficient network and partial correlation coefficient network. Second, we construnct the influence network and mention about its qualitative features. Finally, we examine the indexes which is considering the influence network and useful for individual investors.
本研究では、個人投資家に株式の投資判断に必要な情報 を提示する投資支援システムを作ることを目的とする。そのために、複数のテキスト情報を用いて、銘柄に変動的に影響する要因の分析と、企業の取引企業や事業内容といった基本情報を取得する。
In this paper, we proposed a method that summarizes pdf files of summary of financial statements. Specifically, our method extracts contents of financial statements, causal information and future forecasts from pdf files of summary of financial statements. Then, the summary is generated by connecting them.
It is often said that real estate price information in Japan is less sufficient than that in the US. Especially, we need price estimates of a specific real estate that reflects local characteristics. In this paper, we suggest a local price estimate method using GWR and a web system for publification.
An asset network systemic risk (ANWSER) model is presented to investigate the impact of how shadow banks are intermingled in a financial system on the severity of financial contagion. Particularly, the focus of this study is the impact of the following three representative topologies of an interbank loan network between shadow banks and regulated banks. (1) Random mixing network: shadow banks and regulated banks are intermingled randomly. (2) Asset-correlated mixing network: banks having bigger assets are a regulated bank and other banks are shadow banks. (3) Layered mixing network: banks in a shadow bank layer are connected to banks in a regulated bank layer with some interbank loans.
2014 年3 月にロンドンで行われた国際会議,IEEE Computational Intelligence for Financial Engineering & Eco-nomics (CIFEr) 2014 に参加したので報告する.