CN113254003A

CN113254003A - Editing method and system for quantitative transaction strategy

Info

Publication number: CN113254003A
Application number: CN202110810715.3A
Authority: CN
Inventors: 赵小光; 李建军; 黄彪; 王平平; 谭国苹
Original assignee: Beijing Xinghuo Quantitative Technology Co ltd
Current assignee: Beijing Xinghuo Quantitative Technology Co ltd
Priority date: 2021-07-19
Filing date: 2021-07-19
Publication date: 2021-08-13

Abstract

The invention provides an editing method and an editing system of a quantitative trading strategy, which realize a graphical interaction mode of editing the quantitative trading strategy through a display interface of a graphical strategy editing interface, provide an efficient, convenient and highly extensible editing mode for investors, and the investors can conveniently edit the quantitative trading strategy without simultaneously having programming capability, thereby reducing the writing threshold of the quantitative trading strategy, saving the code writing workload and meeting more user requirements; meanwhile, by providing a complete retest mechanism, the invention enables an investor to pass through the past in a virtual mode to re-verify the effectiveness of a specific quantitative transaction strategy and improve potential benefits.

Description

Editing method and system for quantitative transaction strategy

Technical Field

The invention relates to the technical field of financial information, in particular to an editing method and system for a quantitative transaction strategy.

Background

With the application of quantitative trading strategies in the field of financial market product investment, the establishment of quantitative trading strategies becomes more and more important, and the benefit loss of investment users/teams is directly related to the quality of one quantitative trading strategy.

In the prior art, an investment user/team often uses a programming language (for example, Python language) to write a code of a trading strategy to perform quantitative analysis on data, which requires that the investment user/team needs to be familiar with financial knowledge and have a code success base, the professional requirement threshold is high, and a very small number of investment users/teams (usually trading bulls or opinion leaders) can issue own trading strategies. At present, the data meeting the conditions can be screened in a fixed rule interface manner, for example, some policy indexes are packaged in advance and provided for a user to select in an interface manner, so that the user can create a quantitative transaction policy in a manner of checking the indexes, and the fixed rule interface manner is not high in threshold, but because the user demands are various, some policy indexes packaged in advance by the system are difficult to meet the personalized demands of the user, and the effect is not good.

Disclosure of Invention

The invention aims to provide a method and a system for editing a quantitative trading strategy, which are used for solving the technical problems that the code compiling technology of the conventional quantitative trading strategy is high in threshold, and a fixed rule interface cannot meet the requirements of more users.

In order to achieve the above object, the present invention provides an editing method of a quantified transaction policy, comprising:

step S1: providing a display interface of a graphical strategy editing interface, wherein the display interface is used for providing a display interface of a quantitative trading strategy and an editing interface of elements of the display interface, and executing the step S2;

step S2: performing initial policy editing through the display interface to generate a specific quantification transaction policy, and performing step S3;

step S3: converting the specific quantitative transaction strategy into a specific strategy model, and executing the step S4;

step S4: carrying out retesting on the specific strategy model, judging whether a retesting result meets a first expected requirement, and returning to the step S1 to correct the specific strategy model according to the retesting result when the retesting result does not meet the first expected requirement; and saving the specific strategy model to a transaction strategy library after the return test result meets the first expected requirement.

Optionally, the initial policy editing is performed in the display interface in a click or drag manner.

Optionally, the elements for editing the quantitative trading strategy at least include one or more of trading targets, binning timing and binning timing.

Optionally, the graphical policy editing interface at least includes one or more of a timing relationship, a trade mark, a fluctuation range, a market cycle, a screening start time, a screening end time, and a policy logic relationship.

Optionally, after the step S3 is executed, in a step S4, the specific policy model is retested, and whether a retest result meets a first expected requirement is determined, which specifically includes the following operation steps;

and step S41, carrying out statistical analysis on the specific strategy model by using the historical market quotation data, and judging whether the statistical result meets a first expected requirement.

Optionally, when the specific policy model is tested back, the testing back setting at least includes one or more of a policy name, a policy market, a policy target, a policy condition, a testing back period, a delay, an initial amount, a maximum position, a transaction currency, a lever ratio, a transaction instruction, a stop and loss, and a transaction unit.

Optionally, after step S4, the method further includes:

step S5: judging whether the retest result meets a simulated transaction condition, and returning to the step S1 to correct the specific strategy model again when the retest result does not meet the simulated transaction condition; when the retest result meets the simulated transaction condition, controlling the specific quantitative transaction strategy to execute simulated transaction operation according to preset simulated transaction parameters to obtain a simulated transaction result, and executing step S6;

step S6: judging whether the simulated transaction result meets the real-disk transaction condition, and returning to the step S1 to correct the specific strategy model again when the simulated transaction result does not meet the real-disk transaction condition; and when the simulated transaction result meets the real-disk transaction condition, controlling the quantitative transaction strategy to execute real-disk transaction operation according to preset real-disk transaction parameters.

Optionally, the step of controlling the quantitative transaction policy to execute the real-disk transaction operation includes:

step S61: generating a transaction robot according to the specific strategy model, and executing the step S62;

step S62: and the transaction robot triggers a transaction rule according to the real-time market quotation data, and generates and pushes the transaction signal.

The invention also provides an editing system for the quantified transaction strategy, which comprises the following steps:

the display interface of the graphical strategy editing interface is used for providing a display interface of the quantitative transaction strategy and an editing interface of elements of the display interface, and the initial strategy editing is carried out through the display interface so as to generate the specific quantitative transaction strategy;

a quantitative model platform for converting the specific quantitative transaction strategy into a specific strategy model;

the retest system is used for retesting the specific strategy model, judging whether a retest result meets a first expected requirement or not, and correcting the specific strategy model according to the retest result when the retest result does not meet the first expected requirement; and saving the specific strategy model to a transaction strategy library after the return test result meets the first expected requirement.

The invention also provides computer equipment which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor implements the editing method of the quantitative trading strategy when executing the computer program.

The present invention also provides a computer-readable storage medium storing a computer program for executing the editing method of the quantitative transaction policy.

In the editing method and the editing system of the quantitative trading strategy, provided by the invention, the graphical interaction mode of editing the quantitative trading strategy is realized through the display interface of the graphical strategy editing interface, an efficient, convenient and highly extensible editing mode is provided for investors, the investors can conveniently edit the quantitative trading strategy without simultaneously having financial knowledge and programming capacity, the threshold of compiling the quantitative trading strategy can be reduced, the workload of compiling codes is saved, and more user requirements are met; meanwhile, by providing a complete retest mechanism, the invention enables an investor to pass through the past in a virtual mode to re-verify the effectiveness of a specific quantitative transaction strategy and improve potential benefits.

Drawings

FIG. 1 is a flow chart of a method for editing a quantitative transaction policy according to an embodiment of the present invention;

FIG. 2 is a block diagram of an editing system for quantifying transaction policies according to an embodiment of the present invention;

wherein the reference numerals are:

10-display interface; 20-a quantitative model platform; 30-a retest system; 40-a trading strategy library; 50-historical market quotation database; 60-real-time market quotation database.

Detailed Description

The following describes in more detail embodiments of the present invention with reference to the schematic drawings. The advantages and features of the present invention will become more apparent from the following description. It is to be noted that the drawings are in a very simplified form and are not to precise scale, which is merely for the purpose of facilitating and distinctly claiming the embodiments of the present invention.

Fig. 1 is a flowchart of an editing method of a quantitative transaction policy provided in this embodiment. As shown in fig. 1, the present embodiment provides an editing method for a quantified transaction policy, including:

Specifically, step S1 is executed first, and a display interface of the graphical policy editing interface is provided, where the display interface is used to provide a display interface for quantifying the trading policy and an editing interface for its elements.

In this embodiment, the display interface may include a series of wizard configuration interfaces (a policy configuration interface, a transaction operation interface, a transaction condition configuration interface, a risk control condition configuration interface, and the like) based on data driving, and guides a user to visually configure a quantitative transaction policy, so that an investor can edit and display the quantitative transaction policy conveniently.

Optionally, the elements for editing the quantitative trading strategy at least comprise one or more of trading targets, a binning establishment opportunity and a binning leveling opportunity; the graphical strategy editing interface at least comprises one or more of a time sequence relation, a transaction target, a fluctuation range, a market cycle, a screening starting time, a screening ending time and a strategy logic relation; the quantitative trading strategy may be, but is not limited to, a strategy for quantitative trading of financial market products.

Next, step S2 is executed to perform initial policy editing through the display interface to generate a specific quantified transaction policy.

It should be noted that, for the purpose of implementing a visual editing policy, the present embodiment may encapsulate each policy index for financial market product transaction as a code function, and establish an index function library. The investor can edit the initial strategy in the display interface in a clicking or dragging mode, and actually call a packaged code function in an index function library in a mode of selecting the strategy index so as to generate a specific transaction strategy.

Next, step S3 is executed to convert the specific quantified transaction policy into a specific policy model.

The specific strategy model is a data model which can be identified by a computer, and after the specific transaction strategy is converted into the specific strategy model, the computer can conveniently carry out modeling and analysis by the technologies of statistics, big data, machine learning, deep neural network and the like.

For example, for a strategic method scenario based on k-line graph trend, the data elements and relationships are illustrated as follows: the time sequence relation is as follows: the pull-down lists comprise a first list, a second list, a third list, a fourth list and the like, and are used for identifying the sequential relation of the K line graphs; the trade target is as follows: searching an input box, wherein the content is a stock or futures contract code and is used for identifying a trade object or a trigger condition of reducing/leveling dependence; the market cycle is as follows: a pull-down list, the contents of which are 5 minutes, 30 minutes, 1 hour, 1 day, 1 week, 1 month, 5 days average line deviation degree, 10 days average line deviation degree and the like, for identifying the change period of the K-line graph; the rising and falling amplitude: the input box is used for defining the fluctuation proportion of a target object in a K line graph in a certain period; the screening starting time and the screening ending time are the starting time and the ending time determined by searching the historical data; and (3) rise-fall comparison: the policy logic relationship is as follows: the drop-down list, the content of which is and or, is used to identify the logical relationship of different policy conditions.

In another case, the main content of the initial policy editing described above relates to: including trade mark (stock or futures contract code), fluctuation range (percentage), market cycle (5 minutes, 30 minutes, 1 hour, 1 day, 1 week, 1 month, 5 days average deviation, 10 days average deviation, etc.), screening start time (e.g., 1 month and 1 day of 2021 year), screening end time (e.g., 4 months and 30 days of 2021 year), multi-strategy logical relationship (and, or), time series relationship (first, second, third, fourth), etc.; the edited contents have low threshold for users, are convenient to understand and design, and can be conveniently mapped to basic data units used by a quantitative model after being processed, so that the application threshold of the users is reduced, the complex processing process is solved, and the quantitative modeling is convenient.

In the specific embodiment of the application, a user sets statistical conditions and statistical result constraints; the statistical conditions mainly refer to information settings such as time sequence relation, trade mark, market period, screening starting time, screening ending time, strategy logic relation and the like; in this embodiment, the statistical condition is mapped to a data set 1 and matched in the historical market data, and the triggering frequency meeting the statistical condition and the fluctuation range information of 5 minutes, 15 minutes, 30 minutes, 1 hour, 1 day, 1 week and 1 month after triggering are counted in the period 1 range; and finally, screening the triggering frequency and the fluctuation amplitude information after triggering according to the statistical result constraint set by the user and displaying the information to the user in a scatter diagram and text description mode.

Statistical constraints include: trade mark (stock or futures contract code), market period (5 minutes, 30 minutes, 1 hour, 1 day, 1 week, 1 month), direction of rise and fall (rise or fall is greater than or equal to), magnitude of rise and fall (percentage);

next, in step S41, a statistical analysis is performed on the specific strategy model using the historical market data, and it is determined whether the statistical result satisfies a first expected requirement.

Specifically, the specific strategy model is associated with a historical market quotation database, and then the historical market quotation data in the historical market quotation database is input into the specific strategy model to perform statistical analysis on the specific strategy model, wherein the statistical analysis can be probability calculation, machine learning/deep learning modeling and the like.

And feeding back the statistical result to the investor, and then judging whether the statistical result meets a first expected requirement, wherein the first expected requirement can be an empirical value obtained in practical application or a preset value set by the investor. When the statistical result does not meet the first expected requirement, the investor can edit the initial strategy again on the display interface according to the statistical result and the preference of the investor, so that a new specific transaction strategy is obtained, and the specific strategy model is corrected. When the statistical result meets the first expected requirement, it may be preliminarily determined that the particular policy model meets the requirement.

Next, performing step S4, performing a retest on the specific policy model, determining whether a retest result meets a first expected requirement, and returning to step S1 to correct the specific policy model according to the retest result when the retest result does not meet the first expected requirement; and saving the specific strategy model to a transaction strategy library after the return test result meets the first expected requirement.

Specifically, market quotation data in a retest time period are acquired from a time sequence database or a relational database; and inputting market quotation data in the retest period into the specific strategy model one by one in a mode of message queue, so as to obtain a retest result. The main process of machine learning the information is as follows: mapping risk bearing capacity and expected profit and loss information of a user into a first expected requirement, extracting a discontinuous training data set 1 in a period 1 range in a historical database, modeling and predicting the quotation next to the period 1, comparing the quotation with real quotation data to calculate accuracy, setting an accuracy threshold, and selecting and eliminating models, wherein only the selected model outputs a prediction result; the technical innovation point of the embodiment of the application is that data features are generated through a discontinuous training data set, subsequent continuous data are predicted, models are selected and eliminated through setting of threshold values, and potential prediction errors caused by fixed models are avoided to the maximum extent.

In particular embodiments of the present application, the first contemplated claim set element may be varied; for example, the first expected demand may be a user profitability; or the first expected requirement can be a set expanding expected value and a set retracting expected value, and the discontinuous training data set 1 is extracted within the cycle 1 range; then, the features of the discontinuous data set 1 are screened and extracted, test verification is carried out in the period 2 range, the feature value set is supplemented according to the test verification effect, and iteration is carried out on the following period 3 and period 4 until the effect meets the expected requirements of users.

In this embodiment, when the specific policy model is tested back, the testing device at least includes one or more of a policy name, a policy market, a policy target, a policy condition, a testing back period, a delay, an initial amount, a maximum position, a transaction currency, a lever ratio, a transaction instruction, a stop and loss, and a transaction unit.

It should be understood that after a user writes a specific quantitative transaction strategy, multiple manual operations are often required to perform a back test or a simulated transaction test of the strategy, check the test result, and perform real-disk transaction under the condition of satisfying the result, the whole process is time-consuming and labor-consuming, and the time-consuming specific strategy model test is too long due to the change of market conditions with time, so that the instantaneity of real-disk transaction is affected, and thus, the investment user misses the best opportunity for releasing the specific quantitative transaction strategy, and the investment profit is reduced.

Therefore, in this embodiment, step S5 is executed to save the specific policy model in the transaction policy repository, and a simulated transaction test is automatically executed on the specific policy model.

Step S5 is executed: judging whether the retest result meets a simulated transaction condition, and returning to the step S1 to correct the specific strategy model again when the retest result does not meet the simulated transaction condition; and when the retest result meets the simulated transaction condition, controlling the specific quantitative transaction strategy to execute simulated transaction operation according to preset simulated transaction parameters to obtain a simulated transaction result.

Specifically, whether the retest result meets the simulated transaction condition is monitored, and when the retest result does not meet the simulated transaction condition, the investor can edit the initial strategy again on the display interface according to the retest result and the preference of the investor, so that a new specific transaction strategy is obtained, and the specific strategy model is corrected. And under the condition that the retest result meets the simulated transaction condition, controlling the specific quantitative transaction strategy to execute simulated transaction operation according to preset simulated transaction parameters to obtain a simulated transaction result. Wherein, whether the retest result meets the simulation transaction condition or not is judged; a specific quantitative transaction strategy model of the transaction robot can be perfected only after a plurality of tests are obtained when simulated transaction conditions are met; the simulation transaction condition is that when the simulation test is carried out, the option of the return test setting reflected by the return test result is consistent with the option when the simulation transaction occurs; namely for example: the return test setting comprises return test period setting and delay setting, so that the two items are required to be triggered under the condition of the simulated transaction, and the simulated transaction operation can be executed by a specific quantitative transaction strategy according to preset simulated transaction parameters to obtain a simulated transaction result; obtaining a simulation transaction result, namely a simulation test result, and then executing the following step S6;

further, executing step S6, determining whether the simulated transaction result satisfies the real transaction condition, and returning to step S1 to correct the specific strategy model again when the simulated transaction result does not satisfy the real transaction condition; and when the simulated transaction result meets the real-disk transaction condition, controlling the quantitative transaction strategy to execute real-disk transaction operation according to preset real-disk transaction parameters.

Specifically, whether the simulated transaction result meets the real-market transaction condition is monitored, and when the simulated transaction result does not meet the real-market transaction condition, the investor can edit the initial strategy again on the display interface according to the return test result and the preference of the investor, so that a new specific transaction strategy is obtained, and the specific strategy model is corrected. And when the simulated transaction result meets the real-disk transaction condition, controlling the quantitative transaction strategy to execute real-disk transaction operation according to preset real-disk transaction parameters to obtain a real-disk transaction result. The real disk transaction condition may be a condition that a simulated transaction result is required to be satisfied when the real disk transaction operation is automatically executed. Wherein, step S6 executes: after the simulated transaction result is obtained, the real-disk transaction test should be carried out; the real-disk transaction condition means that the option when the simulated transaction occurs reflected by the simulated transaction result is consistent with the option when the real-time market transaction occurs (only the real-disk transaction can be verified); then, the specific quantitative trading strategy executes the simulated trading operation according to the preset real-disk trading parameters to obtain a real-disk trading result; and pushing the obtained real disk transaction result to an investor or a transaction system.

Preferably, the statistical result, the return test result, the simulated transaction result and the real-disk transaction result can be fed back to the investor, so that the investor can analyze the specific quantitative transaction strategy.

It is understood that the step of controlling the quantitative transaction policy to perform real disk transaction operations includes:

and executing the step S61, and generating the transaction robot according to the specific strategy model.

And step S62 is executed, the trading robot triggers trading rules according to the real-time market quotation data, and generates and pushes the trading signals. The trading signal may be pushed to an investor or a trading system.

Further, whether historical market data or real-time market data are input into the specific strategy model, the market data can be input into the specific strategy model one by one in a message queue mode. The historical market quotation data comprises historical data information such as securities, futures and foreign currencies; the real-time market data comprises real-time information such as securities, futures, foreign currencies and the like.

Fig. 2 is a block diagram illustrating a structure of an editing system for quantifying a transaction policy according to this embodiment. As shown in fig. 2, the system for editing a quantitative transaction policy includes:

the display interface 10 of the graphical strategy editing interface is used for providing a display interface of the quantitative trading strategy and an editing interface of elements of the display interface, and initial strategy editing is carried out through the display interface so as to generate the specific quantitative trading strategy;

a quantitative model platform 20 for converting the specific quantitative transaction strategy into a specific strategy model;

the retest system 30 is configured to retest the specific policy model, determine whether a retest result meets a first expected requirement, and correct the specific policy model according to the retest result when the retest result does not meet the first expected requirement; when the backtesting result meets the first expected requirement, the specific policy model is saved to the transaction policy repository 40.

Optionally, the editing system for quantifying the trading strategy further includes a historical market quotation database 50 and a real-time market quotation database 60, which are used for respectively providing the historical market quotation data and the real-time market quotation data.

The present embodiment further provides a computer device, which is used to solve the technical problems that in the prior art, a code writing technology threshold of a quantitative transaction policy is high, and a fixed rule interface cannot meet the requirements of more users.

An embodiment of the present invention further provides a computer-readable storage medium, where a computer program for executing the editing method of the quantitative transaction policy is stored in the computer-readable storage medium.

In summary, the present embodiment provides an editing method and system for a quantitative transaction policy, which implement a graphical interaction manner for editing a quantitative transaction policy through a display interface of a graphical policy editing interface, and provide an efficient, convenient, and highly extensible editing manner for investors, so that investors can edit a quantitative transaction policy conveniently without having financial knowledge and programming capability at the same time, thereby reducing the threshold for compiling the quantitative transaction policy, saving code compiling workload, and satisfying more user requirements; meanwhile, by providing a complete retest mechanism, the invention enables an investor to pass through the past in a virtual mode to re-verify the effectiveness of a specific quantitative transaction strategy, reduces error cost by history, and improves potential income.

The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A method of compiling a quantified transaction policy, comprising:

2. The method for editing a quantitative transaction policy of claim 1, wherein initial policy editing is performed by clicking or dragging in the display interface.

3. The method of claim 1, wherein the elements for editing the quantitative trading strategy comprise at least one or more of trading target, binning timing and binning timing.

4. The method for editing a quantitative transaction policy of claim 1 or 3, wherein the graphical policy editing interface comprises at least one or more of timing relationship, transaction target, fluctuation amplitude, market cycle, filtering start time, filtering end time and policy logic relationship.

5. The method for editing a quantitative transaction strategy according to claim 1, wherein after step S3, in step S4, the specific strategy model is tested back and whether the test back result satisfies the first expected requirement is determined, which comprises the following steps;

6. The method of claim 1, wherein when the specific policy model is tested back, the testing settings at least include one or more of policy name, policy market, policy target, policy condition, testing back period, delay, initial amount, maximum position, transaction currency, lever ratio, transaction instruction, stop and lose, and transaction unit.

7. The method for editing a quantitative transaction policy according to claim 1, further comprising, after the step S4:

8. The method for editing a quantitative transaction policy of claim 7, wherein the step of controlling the quantitative transaction policy to perform real-disk transaction operations comprises:

9. An authoring system for quantifying trading strategies, comprising:

10. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements a method of compiling a quantitative transaction policy according to any one of claims 1 to 8 when executing the computer program.