Full-automatic quantification transaction system based on multi-dimension screening and dynamic batch transaction

Abstract

This invention provides a fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading, belonging to the field of trading information management technology. Its features include a data acquisition module, a low-volume identification module, a dynamic batch trading execution module, and a multi-risk management module. The data acquisition module collects, organizes, and uploads information; the low-volume identification module filters and continuously monitors data content; the dynamic batch trading execution module generates an execution plan based on the data and generates an initial position order based on dynamic data changes, while continuously performing purchase transactions in batches; the multi-risk management module receives trading operation information from the dynamic batch trading execution module and determines whether to stop or continue operation based on a preset risk threshold. The advantages of this invention are: it can perform accurate assessment through multi-dimensional data and derive flexible trading strategies to reduce overall risk, while also enabling rapid and accurate interception to reduce losses.

Description

Technical Field

[0001] This invention relates to the field of transaction information management technology, and in particular to a fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading. Background Technology

[0002] With the rapid development of fintech, more and more AI and big data technologies are being applied to the financial market to process stock market data more simply and quickly, significantly reducing the pressure of manual processing and interaction, while ensuring the accuracy of data processing.

[0003] However, existing technologies still suffer from problems such as a relatively singular dimension and insufficient comprehensive judgment in the selection of investment targets. Most systems rely solely on technical indicators derived from historical prices and trading volumes for stock selection, lacking a systematic assessment of the company's fundamental quality and valuation level.

[0004] In terms of trading strategy execution, existing strategies are too rigid and have poor overall adaptability. They cannot respond dynamically to timely adjustments in trading duration, resulting in increased transaction costs and low capital utilization efficiency.

[0005] In terms of risk control, the existing system's risk control module is a remedial operation after the transaction is completed. It is slow to respond and isolated, and cannot obtain risk information in a timely manner for timely management.

[0006] Therefore, there is an urgent need for a system that can evaluate trading strategies from multiple dimensions and flexibly execute them, and fully and timely manage overall risks to solve the above problems. Summary of the Invention

[0007] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0008] This fully automated quantitative trading system, based on multi-dimensional screening and dynamic batch trading, includes a data acquisition module, a low-volume identification module, a dynamic batch trading execution module, and a multi-risk management module.

[0009] The data acquisition module collects transaction information via the Internet, organizes and packages information from multiple channels, and uploads it to the low-volume identification module.

[0010] After filtering and identifying the data content, the low-volume identification module determines whether the current stock price data is in a low-volume state based on historical data content, and sends an instruction to the data acquisition module for continuous monitoring.

[0011] After obtaining the data from the low-volume identification module, the dynamic batch transaction execution module obtains the execution plan data and generates the first position order based on the dynamic data changes, while continuously carrying out purchase transactions in batches.

[0012] The multi-risk management module receives batch purchase transaction operation information from the dynamic batch transaction execution module and analyzes the transaction amount. If the execution plan data exceeds the preset risk threshold, the subsequent plan execution will be automatically stopped. If it meets the preset risk threshold, the plan will be executed and the database information will be updated.

[0013] As an improvement, the data acquisition module also includes a cleaning module. The cleaning module integrates the data content acquired by the data acquisition module based on website information, stock price information, and relevant policy information, and eliminates irrelevant data. The cleaning module uses preset tag information to identify and retain targeted data features.

[0014] As an improvement, a multi-dimensional filtering module is provided between the data acquisition module and the low-volume recognition module. The multi-dimensional filtering module identifies the following based on preset label conditions:

[0015] Fundamentals: Net profit growth rate over the past three months ≥ 10%;

[0016] Valuation: The price-to-earnings ratio is 80% lower than the industry average.

[0017] Volatility: Targets that pass the screening are recorded as "high-quality targets".

[0018] As an improvement, the low-volume identification module combines high-quality target data to identify the timing of the entry and determine whether it is within a dynamically calculated relatively low range. The calculation steps of the relatively low range include: Step 1, defining an initial price range based on the average price of the target over the past 60 days.

[0019] Step 2: Obtain the industry volatility and market benchmark volatility of the target, and dynamically lower the initial price range lower limit according to the formula "Adjusted lower limit of the range = Initial lower limit of the range - (Industry volatility - Benchmark volatility) / Benchmark volatility * Adjustment coefficient", where the adjustment coefficient is a preset percentage;

[0020] If the low-volume recognition module fails to identify the corresponding data signal, it will send the information back to the multi-dimensional filtering module for re-recognition.

[0021] As an improvement, the buying steps of the dynamic batch trading execution module include: Step 1, initial position building: when a buy signal is triggered, buy 30% of the preset total position;

[0022] Step 2, Averaging Down: After buying, if the price of the target stock falls back to the lower limit of the dynamically calculated relative low range, then add 20% of the preset total position.

[0023] Step 3, Breakout Add-on: If the target breaks through the highest price in the past 10 days with a large volume, then add another 20% of the preset total position.

[0024] As an improvement, the dynamic batch selling strategy steps of the dynamic batch trading execution module include: First profit-taking: when the price of the target rises to 110% of the average price of the past 60 days, sell 40% of the held position; Second profit-taking: when the price of the target rises to 120% of the average price of the past 60 days, sell 30% of the held position; Trailing stop loss: when the price of the target falls below the moving average line within 10 days, sell all remaining positions.

[0025] As an improvement, the multi-risk management module will conduct real-time review and verification of the target data information and the account-level data. If the received order violates the risk control rules, it will be immediately intercepted and an alarm will be triggered. The risk control rules can be updated by manually inputting the corresponding parameters through the terminal.

[0026] As an improvement, after the data information of the multi-risk management module passes the risk control review, the order information is sent to the exchange for execution through the brokerage API. After execution, the feedback data information is immediately sent back to the database. The database also includes a status update module. After the status update module identifies the order data update information, it uploads it to the multi-dimensional filtering module to determine whether the data information after the transaction meets the further transaction requirements.

[0027] The beneficial effects of this invention are: by setting up a data acquisition module and a multi-dimensional filtering module, it realizes multi-dimensional parallel filtering of fundamentals, valuation, and volatility, making the stock selection data more comprehensive and stable;

[0028] By setting up a low-volume recognition module and a dynamic batch transaction execution module, dynamic data processing can be achieved, avoiding the uncontrollability caused by a one-time transaction. Multiple transactions can reduce losses from stock market fluctuations and amplify overall returns.

[0029] By setting up multiple risk control modules, multiple risk control rules can be obtained and intercepted at the strategy, transaction order and execution stages. Combined with multi-dimensional screening modules, risk information can be processed in a timely manner, improving system stability and reducing risks. Attached Figure Description

[0030] Figure 1 This is a flowchart of the fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading according to the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0032] It should be noted that the terms "first" and "second" in this application are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0033] like Figure 1 As shown, the fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading includes a data acquisition module, a low-volume identification module, a dynamic batch trading execution module, and a multi-risk management module.

[0034] The data acquisition module collects transaction information via the internet, organizes and packages information from multiple channels, and uploads it to the low-volume identification module. After filtering and identifying the data, the low-volume identification module determines whether the current stock price is in a low-volume state based on historical data and sends instructions to the data acquisition module for continuous monitoring. The dynamic batch transaction execution module obtains the data from the low-volume identification module, generates execution plan data, and generates an initial position order based on dynamic data changes, while continuously performing batch purchase transactions. The multi-risk management module receives batch purchase transaction information from the dynamic batch transaction execution module and analyzes the transaction amount. If the execution plan data exceeds a preset risk threshold, subsequent plan execution is automatically stopped; otherwise, the plan is executed, and the database information is updated. The data acquisition module can analyze financial news and social media sentiment through a natural language processing module to generate sentiment factors, assessing the impact on the stock market and strategy tendencies from social data sources. The data acquisition module accesses different data sources through multiple interfaces, expanding the overall data source information, further improving data diversity, and ensuring the overall strategy's flexibility and accuracy.

[0035] The data acquisition module also includes a cleaning module. This cleaning module integrates the data acquired by the data acquisition module based on website information, stock price information, and relevant policy information, eliminating irrelevant data. The cleaning module uses preset tag information to identify and retain targeted data features. A multi-dimensional filtering module is located between the data acquisition module and the low-volume identification module. This multi-dimensional filtering module identifies the following based on preset tag conditions: Fundamentals: Net profit growth rate ≥ 10% in the past three months; Valuation: Price-to-Earnings ratio lower than 80% of the average price-to-earnings ratio of the indicated industry; Volatility: Stocks that pass the filtering are recorded as "high-quality stocks". These tag conditions can be further analyzed and confirmed based on additional data requirements, comprehensively considering the impact of the market environment and policy environment on the stock market and fully filtering out "high-quality stocks".

[0036] The low-volume surge identification module combines high-quality target data to identify entry opportunities and determine whether the target is within a dynamically calculated relatively low price range. The calculation steps for this relatively low price range include: Step 1, defining an initial price range based on the target's 60-day average price; Step 2, obtaining the target's industry volatility and market benchmark volatility, and dynamically lowering the initial price range lower limit according to the formula "Adjusted lower limit = Initial lower limit - (Industry volatility - Benchmark volatility) / Benchmark volatility * Adjustment coefficient," where the adjustment coefficient is a preset percentage. If the low-volume surge identification module fails to identify the corresponding data signal, it sends the information back to the multi-dimensional filtering module for re-identification. After the data calculation is complete, the target can be filtered a second time, and the system logic determines the buying opportunity and generates specific trading signals. These trading signals can retrieve corresponding execution plans from the database and be finely adjusted.

[0037] The buying steps of the dynamic batch trading execution module include: Step 1, initial position building: When a buy signal is triggered, buy 30% of the preset total position; Step 2, pullback averaging down: After buying, if the price of the target falls back to the lower limit of the dynamically calculated relative low range, then buy an additional 20% of the preset total position; Step 3, breakout averaging down: If the target breaks through the highest price of the past 10 days with increased volume, then buy another 20% of the preset total position. The dynamic batch selling strategy steps of the dynamic batch trading execution module include: First profit-taking: When the price of the target rises to 110% of the past 60-day moving average, sell 40% of the held position; Second profit-taking: When the price of the target rises to 120% of the past 60-day moving average, sell 30% of the held position; Trailing stop-loss: When the price of the target falls below the 10-day moving average, sell all remaining positions. The multi-risk management module performs real-time review and verification of the target data and account-level data. If a received order violates risk control rules, it will be immediately blocked and an alarm will be triggered. The risk control rules can be manually updated by inputting corresponding parameters through the terminal. After the data information of the multi-risk management module passes the risk control review, the order information is sent to the exchange for execution via the brokerage API. After execution, the feedback data information is immediately sent back to the database. The database also includes a status update module. After identifying the order data update information, the status update module uploads it to the multi-dimensional filtering module to determine whether the post-trade data information meets the requirements for further trading. At this time, the virtual position and account status will be updated and identified synchronously to determine whether the current status meets the trading logic status and input the corresponding conditions for the next round of data filtering and trading decisions, realizing a sustainable self-iterative trading closed loop.

[0038] When used, the overall method includes the following steps: Step 1: Real-time collection and cleaning of market data and fundamental data of all financial instruments in the market; Step 2: Based on parallel computing technology, apply multi-dimensional quantitative screening to the cleaned data, including at least fundamental conditions, valuation conditions, and volatility conditions, and mark the instruments that meet all screening conditions as high-quality instruments; Step 3: Identify the buying opportunity for high-quality instruments, determine whether their current price is in a dynamically calculated relatively low range, and whether it meets the volume trigger condition. If both conditions are met, a buy signal is generated; Step 4: In response to the buy signal, an initial position order is automatically generated according to a dynamic batch buying strategy; and after the order is executed, the price of the instrument is continuously monitored. If the price meets the pullback averaging down condition or the breakout adding condition, a corresponding additional order is automatically generated; Step 5: During the holding period, the price of the instrument is continuously monitored. If the price touches the tiered profit-taking condition or the trailing stop-loss condition, a corresponding sell order is automatically generated according to a dynamic batch selling strategy.

[0039] During use, once the target is bought and sold, and related transaction orders are completed, the relevant information will be sent to the multi-layer risk control module for real-time compliance verification. The multi-layer risk control module will at least implement target-level position concentration control and account-level overall net value stop-loss line control. If the order passes the verification, it will be sent to the exchange for execution; if it fails, it will be automatically intercepted and an alarm will be triggered.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading, characterized in that: It includes a data acquisition module, a low-volume identification module, a dynamic batch transaction execution module, and a multi-risk management module. The data acquisition module collects transaction information via the Internet, organizes and packages information from multiple channels, and uploads it to the low-volume identification module. After filtering and identifying the data content, the low-volume identification module determines whether the current stock price data is in a low-volume state based on historical data content, and sends an instruction to the data acquisition module for continuous monitoring. After obtaining the data from the low-volume identification module, the dynamic batch transaction execution module obtains the execution plan data and generates the first position order based on the dynamic data changes, while continuously carrying out purchase transactions in batches. The multi-risk management module receives batch purchase transaction operation information from the dynamic batch transaction execution module and analyzes the transaction amount. If the execution plan data exceeds the preset risk threshold, the subsequent plan execution will be automatically stopped. If it meets the preset risk threshold, the plan will be executed and the database information will be updated.

2. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, The data acquisition module also includes a cleaning module. The cleaning module integrates the data acquired by the data acquisition module based on website information, stock price information, and relevant policy information, and eliminates irrelevant data. The cleaning module uses preset tag information to identify and retain targeted data features.

3. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, A multi-dimensional filtering module is provided between the data acquisition module and the low-volume recognition module. The multi-dimensional filtering module identifies the following based on preset label conditions: Fundamentals: Net profit growth rate over the past three months ≥ 10%; Valuation: The price-to-earnings ratio is 80% lower than the industry average. Volatility: Targets that pass the screening are recorded as "high-quality targets".

4. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 3, characterized in that, The low-volume identification module combines high-quality target data to identify the timing of the entry and determine whether it is within a dynamically calculated relatively low range. The calculation steps of the relatively low range include: Step 1: Define an initial price range based on the average price of the target over the past 60 days. Step 2: Obtain the industry volatility and market benchmark volatility of the target, and dynamically lower the initial price range lower limit according to the formula "Adjusted lower limit of the range = Initial lower limit of the range - (Industry volatility - Benchmark volatility) / Benchmark volatility * Adjustment coefficient", where the adjustment coefficient is a preset percentage; If the low-volume recognition module fails to identify the corresponding data signal, it will send the information back to the multi-dimensional filtering module for re-recognition.

5. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, The buying steps of the dynamic batch trading execution module include: Step 1, initial position building: when a buy signal is triggered, buy 30% of the preset total position; Step 2, Averaging Down: After buying, if the price of the target stock falls back to the lower limit of the dynamically calculated relative low range, then add 20% of the preset total position. Step 3, Breakout Add-on: If the target breaks through the highest price in the past 10 days with a large volume, then add another 20% of the preset total position.

6. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, The dynamic batch selling strategy steps of the dynamic batch trading execution module include: First profit-taking: when the price of the target rises to 110% of the average price of the past 60 days, sell 40% of the held position; Second profit-taking: when the price of the target rises to 120% of the average price of the past 60 days, sell 30% of the held position; Trailing stop loss: when the price of the target falls below the 10-day moving average, sell all remaining positions.

7. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, The multi-risk management module performs real-time review and verification of the target data information and account-level data. If the received order violates the risk control rules, it will be immediately intercepted and an alarm will be triggered. The risk control rules can be updated by manually inputting the corresponding parameters through the terminal.

8. The fully automated quantitative trading system based on multi-dimensional screening and dynamic batch trading as described in claim 1, characterized in that, After the data from the multi-risk management module passes the risk control review, the order information is sent to the exchange for execution via the brokerage API. After execution, the feedback data is immediately sent back to the database. The database also includes a status update module. After the status update module identifies the order data update information, it uploads it to the multi-dimensional filtering module to determine whether the post-trade data information meets the requirements for further trading.

Images