Recently, Transwarp Technology held the 2022 Spring New Product Release Week online, bringing four major topics: digital base, digital transformation, domestic substitution, data security and circulation, and more than 60 speeches. Customers and partners from various fields such as finance, energy, transportation, government affairs, universities, and operators shared their experiences in digital transformation. Yang Heguo, the person in charge of technology supervision of Zhengzhou Commodity Exchange, combined his exploration and experience in the field of financial technology, and brought a keynote speech on "Integration of Digital Intelligence and Digital Transformation of Exchanges".
In recent years, the country has attached great importance to the research and application of big data, artificial intelligence and other technologies, and has also clarified the national big data strategy, using big data as a basic strategic resource. As an exchange under the CSRC system, Zhengzhou Commodity Exchange (hereinafter referred to as ZCE) is promoting the development and application of digital resources, as well as transformation and upgrading.
At the same time, since 2017, the China Securities Regulatory Commission has also released a number of application construction plans or plans for big data and artificial intelligence technology. In the field of banking, the People's Bank of China has also issued a number of financial technology application routes and development plans. Especially after the China Securities Regulatory Commission proposed the overall construction plan of regulatory technology in 2018, it also clarified five basic data analysis capabilities and 32 regulatory business analysis scenarios. In September 2021, following the country's "14th Five-Year Plan" strategic plan, the China Securities Regulatory Commission also proposed the "14th Five-Year Plan" for the technological development of the securities and futures industry. , to help China's economy shift from high-speed growth to high-smart development.
TDH, the big data basic platform of Transwarp Technology, helps ZCE's digital transformation
Looking at the development trend of the financial industry, foreign exchanges have changed from a single asset trading center to a global development center, and have gradually undergone digital transformation from core business to external services. Domestic futures exchanges are currently in the early stages of big data application supervision and development, and have also built multiple application scenarios. In general, ZCE not only continues to carry out research and construction of financial technology, but also uses technology to empower and dig deep into the value of data to prevent financial risks, serve the real economy, and further enhance market supervision capabilities and market service levels. It is promoting the high-quality development of the exchange.
Zhengzhou Commercial Exchange has established a financial technology strategic positioning and development goals. From the strategic positioning, it proposes to enhance the ability of technology to support business development, promote the deep integration of business and technology, and ultimately promote the digital transformation of the exchange. To promote the digital transformation of the exchange, it is necessary to improve three aspects: technology supervision capability, market service capability and business operation efficiency.
Since 2017, ZCE has been exploring and researching big data, artificial intelligence and other technologies: in 2018, it made a POC of a big data platform, and finally chose TDH, a basic big data platform of Transwarp Technology with independent property rights in China. And completed the migration of traditional data warehouses in 2018, and built some big data applications; from 2018 to 2020, developed a number of applications based on big data platforms, and won relevant awards from the China Securities Regulatory Commission and the People's Bank of China; in 2021, Propose a financial technology development plan, strengthen the construction of the data center, and build an overall big data platform.
ZCE’s traditional data warehouse architecture is a simple model with poor timeliness and storage space, especially the cost of storage space expansion is also very high. The entire process is adjusted from data access to application scenarios, realizing a multi-heterogeneous architecture and completely replacing the traditional data warehouse architecture. During the four or five years of using Xinghuan technology products, ZCE has also discovered some advantages of TDH products, such as paying attention to auditing, especially the smooth migration of business, which can be done without too much development and workload. Some codes of Oracle were migrated before. Moreover, it also has a machine learning platform, supports multiple languages, and also provides a lot of help for data analysis.
In the future, ZCE will continue to improve its data operation and service capabilities to form an ecological data platform.
Build an AI prediction model to improve the level of intelligent decision-making
In addition, ZCE has also jointly explored three representative projects with Transwarp Technology in terms of AI, namely risk control measures to assist decision-making, abnormal transaction identification, and hedging quota approval.
The futures market mainly connects the real economy and the financial market, effectively making up for the shortcomings of the spot market, and playing an important role in stabilizing and promoting the development of the market economy. In the system and policy of the exchange, measures such as margin, price limit, and handling fee are important means and core means of market risk control and adjustment. How to adjust margin, price change fee, handling fee, and how to effectively control market risk? This has brought great difficulties to ZCE.
Based on this, ZCE has carried out in-depth cooperation with Transwarp Technology to explore the means of using historical data and historical cases to assist in risk control and adjustment, which uses a large amount of transaction flow data and historical parameters, as well as machine learning engines and rule engines. engine. Among them, the rule engine relies on expert experience, and can adjust some adjustment rules based on the historical experience of adjusting various risk control operations; the machine learning engine uses the effect analysis of previous adjustments to form a weighted prediction sequence, and at the same time combines TDH's TensorFlow framework to build, A closed loop is formed from forecast results, visual charts, and automated reports. Finally, before the policy is issued, the potential impact of risk control policies and measures on the futures market can be assessed, and the formulation of trading rules and measures can be assisted, making the policy more prudent, reasonable and effective.
The second case is abnormal transaction identification. In recent years, the trading volume of the futures market has gradually increased, and the participants are intricate. Foreign traders have also been introduced at home and abroad, especially programmatic trading. Its behavioral characteristics are difficult to capture because the trading volume and order volume are very large. How to identify Whether a client's trading behavior is abnormal or aggressive behavior will become very important. To this end, ZCE and Transwarp Technology have also done research in this area, using Transwarp Technology's big data platform TDH and intelligent analysis tool Sophon to build a reverse reinforcement learning model, combined with customer transactions, orders, profits and other transaction characteristics , Reversely predict the potential trading behavior risks in the market, build customer models through risks, and realize more intelligent and scientific precise policy implementation.
The third application scenario is the quota approval for hedging. The most critical part in the approval of hedging quota is how to approve the hedging quota for the customer. The hedging quota is similar to the credit card approval quota, and how much quota is approved to the customer will also determine the subsequent business situation of the customer. The traditional approval mode is that the staff conduct a large amount of historical analysis and historical regression after receiving the submitted materials to see whether the proposed approval amount for the customer is reasonable and scientific, and the work efficiency will be very low. By exploring and using historical data for simulation analysis and automatically generating recommended quotas, hedging personnel can configure calculation rules and generate reports for different varieties, different general months, and adjacent months of hedging according to actual needs. In this way, quota recommendations will be generated within 3 minutes after members submit their submissions, which greatly improves the approval efficiency of hedging quotas.
The current financial technology in the futures market is still in the initial stage of exploration, and there are also deficiencies in financial technology data accumulation and governance. With the acceleration of financial innovation, the China Securities Regulatory Commission has also strengthened the requirements for technology supervision, requiring the industry to accelerate digital transformation. ZCE will continue to deepen financial technology research, strengthen technological innovation and data applications, and improve regulatory efficiency and service levels.
In the next step, ZCE will focus on four aspects for planning: one is to promote the innovation of big data services, that is, to empower businesses with rapid innovation and low-cost trial and error capabilities, and to promote the innovation of big data services. The second is to reduce the cost of informatization construction and operation and maintenance, increase investment in cloud computing, realize unified management and sharing, and greatly reduce the cost of informatization operation and maintenance. The third is to continue to improve the data governance system, unify the data storage and processing platform, upgrade and optimize the big data platform architecture, and improve data analysis capabilities and compliance supervision capabilities. The fourth is to promote the development of the exchange's data ecology. It will integrate data applications, AI models and other aspects to form a closed loop of construction and promote the ecological development of the exchange.