Use black box to do automatic transaction

There are many researches and developments on automatic trading using AI technology such as deep learning represented by ultra-high speed trading. However, automatic trading can not cope with sudden events, increasing the risk of adversely affecting the market. In particular, many automatic trading systems do not show the reason or evidence for investment behavior, the readability of processes and results is low, and they are black boxed, resulting in high voices of anxiety. Therefore, the purpose of this research is to establish a foundation for elucidating the influence of improving the legibility of investment on the investment decision by specifying the investment evidence and reasons.

Research Contents

Comparison of investment products

In this research, one of the topic is to establish a factor analysis base of investment products to clarify the influence of events on the market. We analyze the factors quantitatively and globally using a dynamic spatial model. In addition, we propose methods to efficiently analyze across various investment data such as investment history, market information, investment reports, news reports and so on.

Characteristic analysis of investors

By utilizing financial engineering knowledge such as risk management and portfolio theory, we perform comparisons among investors and time-series analysis of investment history to extract characteristics such as areas in which investors are good and consistency of investment behavior. By doing so, we will build an investor's feature analysis basis that will clarify the characteristics (when and how to trade) of investors' investment behavior.

Investment support system

We compare investment products, analyze the characteristics of investors, and use those data to build an investment system. The support system provides a basis for identifying investment evidence and reasons, and for elucidating how readability influences decision-making. By doing so, it aims to relieve the unrest of the consumer, who is the original investor, and to lead to an increase in investment.



  • 1. Zimao Liu, Qiang Ma:Unsupervised Method for Discovering Expert Traders on Social Trading Services. BigComp 2019: 1-8
  • 2. Woonyeol Lee, Qiang Ma: Discovering Expert Traders on Social Trading Services. JACIII 22(2): 224-235 (2018)
  • 3. Makoto Kirihata, Qiang Ma: Global Analysis of Factors by Considering Trends to Investment Support. DEXA (1) 2018: 119-133
  • 4. Nobuaki Onishi and Qiang Ma: Factor Analysis of Investment Trust Products by Using Monthly Reports and News Articles, ICDIM2017:32-37,2017.
  • 5. Satoshi Baba, and Qiang Ma: Analyzing Relationships of Listed Companies with Stock Prices and News Articles. DEXA (2) 2016: 27-34,2016
  • 6. Yuki Awano, Qiang Ma, and Masatoshi Yoshikawa: Causal Analysis for Supporting Users' Understanding of Investment Trusts. iiWAS 2014: 524-528,2014
  • 7. 馬強:投資支援のための関係マイニング技術の開発,TELECOM FRONTIER(95),2017
  • •. 科学研究補助金(基盤(B)),「エビデンスベースの投資支援に向けたエンティティ指向投資ビッグデータ分析基盤の構築」,2019-
  • •. 京都大学教育研究振興財団 研究助成 「エビデンスベースの投資に向けたエンティティ指向投資データ分析基盤の開発」,2018