(What) our theme
"Financial risk monitoring and management based on the user's portrait"
(Who) our participants
Liu Qian, Li Jian, Hang Lung Chan, Gao origin, Teng Lin, Zhou Yue, Zhao Yun, Wei getting Jun, Shen a chip, Lv Xinran
Model and build: Zhan Hang Lung (person in charge), Liu Qian, Lin Teng, Gao origin, Zhou music
machine learning algorithm: Lin Teng (person in charge), Wei gradually Jun, Zhao Yun
finance theory and financial innovation: Liu Qian (person in charge), Lv Xinran, Hang Lung Chan, Gao origins
back-end technology integration: Li Jian (person in charge), a chip Shen, Zhao Yun, Wei Jun gradually
reception technology integration: Zhou Yue (person in charge), Li Jian, Shen a chip, Lv Xinran
We interspersed with each other in various fields, according to their interests and energy to assist each other, there is no limit.
(Domain) knowledge background and significance
Financial risks include: market risk, credit risk, operational risk, liquidity risk, reputation and strategic risk (defined bar III)
The main significance of research on the subject is active with the current trading center of this, the demand for foreign currency market monitoring of business development, innovation and insights put forward the demand for market monitoring and surveillance methods, to provide a more valuable market for the trading center of financial risk and regulatory bodies identification and monitoring tools to eventually be able to timely and accurate identification of risk factors in financial markets, forecasting risk events.
(Thought) thinking
Unusually large field of financial risks, we can be very flexible in this issue, we can retreat into, can be advanced AI technology, expand innovation in business. But when we started this subject, the focus is to focus on the following main issues: do we mind is Why this subject? What we can provide value to our customers? Please think about this problem continues. And strive to focus on problem areas after June 11.
Thinking 1: Can the transactions motive user modeling, modeling how
Reflection 2: how to deal risks visual display, express and show what form
Reflections: Do we want to focus on their own issues trading risk? Or do we need to stick to this transaction risk children, can diverge farther?
Thinking 4: rule-based risk identification, risk regulation as a business input, the result of a rule-based run out of data, easy to quickly identify business risks as much as possible, this tool can output what he wanted
Voting Topic 1: Our task is to target users: the trading center for the central bank and financial risk-based surveillance, risk identification that provide value added services, supplemented by trading institutions.
Unanimously approved.
Voting Topic 2: Our team has determined whether a merger and market issues directly
Unanimously, the direction object is inconsistent, it is difficult to merge.
Ideas 5: according to the user's transaction behavior data, in the bond market, foreign exchange market, the stock market, trading serial occurred in the gold market, trading portfolio, the trading data to identify bodies transaction (related transactions) behavior derived Financial risk
6 ideas: we can use the technology include: machine learning, mainly RNN neural network algorithm, the traditional statistical methods principal component analysis (PCA), VaR and other financial risk analysis, according to transaction data by targeting the "replay" business logic to identify risks, and ultimately build real-time to identify financial risks recognition engine.
7 thinking: If we think about risk in terms of organization, we need to know the trading strategies institutions need to know the risk appetite of the organization, and the need to know the relationship between credit institutions.
(Work) task assignments
- Jun Wei gradually issued PPT Han Shuo provide financial risk monitoring products
- No. 6 June for the second topic of discussion, let us extensive data collection, focusing on the content of financial risks, as well as methods of analysis and presentation.
(Why) questions
- How do we get the data, which is a problem. We need to get the kind of data to support us to build analytical models and analytical framework.
- What we focus on financial risk areas and key breakthroughs in the end it is? It must be further refined.
The main battlefield of research (Where)
- Building 24202 is our regular meeting room
- Create a collaborative work environment (for example, wps coordination office, the proper way collaboration or other effective system tools, this subject to be voted)
- You need to buy Ali cloud server (this issue to be voted)
(How) How do we complete our task
- We need what kind of modeling tools, write your own program, what kind of language and tools to write programs if they use?
- They receive according to their interest and ability to split tasks, on-time delivery of data collection, form documents, programs and PPT.
- Polymerization and organize research materials to form the final Presentation.
Reproduced in: https: //www.jianshu.com/p/200312a2185d