Artificial Financial Market to analyze algorithm trade and institution

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K01195
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Social systems engineering/Safety system
• Research Institution: Osaka City University
• Principal Investigator: Nakajima Yoshihiro 大阪市立大学, 大学院経済学研究科, 教授 (40336798)
• Co-Investigator(Kenkyū-buntansha): 森 直樹 大阪府立大学, 工学(系)研究科(研究院), 准教授 (90295717)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 人工市場 / 高頻度アルゴリズム取引 / 人工知能 / 機械学習 / 深層学習

Outline of Final Research Achievements

We develop analytical tools of order flow and compare between data of 2007 and 2012. Because TSE Arrowhead started 2010, we can find effects of HFT. At the first step, we count the number of events (including market orders, limit orders and cancel). The number of event increase 5 times from 2007 to 2012, but half of brands decrease the number of events. On the other hand, 22% of the brands increase more than ten times. It suggests that HFT trades mainly in a part of market. Next, We try to use Machine Learning especially Deep learning to apply market analysis. We developed four kinds of models, and observe their profit. Then we find there is no strongest model in these four models.

Free Research Field

計算機経済学