AI-based risk control strategy analysis method and system

Abstract

The present invention relates to an AI-based risk control strategy analysis method and system, in which a risk control strategy matching vector is determined based on an optimized response vector binary, and a risk control strategy response focus vector is determined through response focus processing. Therefore, the risk control strategy matching vector and the risk control strategy response focus vector can improve the detail characterization capability of the response behavior characteristics. In this way, when matching risk control strategy label data through the risk control strategy matching vector and the risk control strategy response focus vector, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed can be accurately and reliably determined, thereby achieving accurate label classification of the risk control strategy execution text to be analyzed, providing convenience for subsequent rapid and accurate calling and retrieval of the risk control strategy execution text to be analyzed, avoiding traversal of the entire text database, and reducing unnecessary resource overhead.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to an AI-based risk control strategy analysis method and system. Background Art

[0002] Risk control was initially applied in the financial sector to provide a reference for business security. With the development of digital businesses, risk control has gradually been applied to other fields, such as digital office and enterprise services. Risk control strategies play a crucial role in this process. Faced with an increasingly complex business environment, the types and number of risk control strategies have increased dramatically. To achieve efficient and accurate risk control strategy management and analysis, it is often necessary to categorize these strategies. However, traditional risk control strategy analysis techniques suffer from low accuracy and efficiency. Summary of the Invention

[0003] In a first aspect, an embodiment of the present invention provides an AI-based risk control strategy analysis method, which is applied to a risk control strategy analysis system. The method includes:

[0004] Obtaining a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed, and obtaining second original risk control policy response vectors corresponding to at least two target risk control policy execution text sets, respectively; wherein the risk control policy label data corresponding to the at least two target risk control policy execution text sets are different;

[0005] Performing knowledge optimization on basic response vector tuples corresponding to at least two second original risk control strategy response vectors to obtain optimized response vector tuples corresponding to each basic response vector tuple; wherein a basic response vector tuple includes the first original risk control strategy response vector and a second original risk control strategy response vector;

[0006] Determining, using at least two optimized response vector binary pairs, a risk control strategy matching vector that reflects the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed; performing response focusing processing on the first original risk control strategy response vector to obtain a risk control strategy response focusing vector corresponding to the risk control strategy execution text to be analyzed;

[0007] According to the risk control policy matching vector and the risk control policy response focus vector, the risk control policy label data corresponding to the risk control policy execution text to be analyzed is determined from at least two risk control policy label data.

[0008] In some example embodiments, performing knowledge optimization on the basic response vector tuples corresponding to the at least two second original risk control strategy response vectors to obtain an optimized response vector tuple corresponding to each basic response vector tuple includes:

[0009] Obtain a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a knowledge feature processing module; the at least two target risk control strategy execution text sets include a target risk control strategy execution text set target set1_x, and the at least two second original risk control strategy response vectors include a second original risk control strategy response vector target vector_x corresponding to the target risk control strategy execution text set target set1_x, where x is a positive integer and is not greater than the total number of corresponding at least two target risk control strategy execution text sets;

[0010] Loading the basic response vector double vector_x consisting of the second original risk control strategy response vector target vector_x and the first original risk control strategy response vector into the knowledge feature processing module;

[0011] Through the knowledge feature processing module, the basic response vector binary double vector_x is subjected to knowledge optimization to obtain a second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and a first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed; wherein, the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x is associated with the risk control strategy execution text to be analyzed; the first optimized risk control strategy response vector majorization vector_x is associated with the target risk control strategy execution text set target set1_x;

[0012] The second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x and the first optimized risk control strategy response vector majorization vector_x are combined to form an optimized response vector double tuple corresponding to the basic response vector double vector_x.

[0013] In some example embodiments, the knowledge feature processing module includes a first knowledge feature processing layer and a second knowledge feature processing layer; the knowledge feature processing module performs knowledge optimization on the basic response vector doublevector_x to obtain a second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x and a first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed, including:

[0014] In the first knowledge feature processing layer, the policy triggering behavior vector in the first original risk control policy response vector is updated using the second original risk control policy response vector targetvector_x to obtain a first optimized risk control policy response vector majorization vector_x corresponding to the risk control policy execution text to be analyzed;

[0015] In the second knowledge feature processing layer, the policy triggering behavior vector in the second original risk control policy response vector target vector_x is updated through the first original risk control policy response vector to obtain the second optimized risk control policy response vector corresponding to the target risk control policy execution text set target set1_x.

[0016] In some example embodiments, the first knowledge feature processing layer includes a group decision unit and an optimization decision unit; the second original risk control strategy response vector target vector_x includes at least two second response event sub-vectors; the at least two second response event sub-vectors include a second response event sub-vector target part_y; the first original risk control strategy response vector includes at least two first response event sub-vectors; the at least two first response event sub-vectors include a first response event sub-vector pending part_y, where y is a positive integer and y is not greater than the total number corresponding to the at least two first response event sub-vectors;

[0017] In the first knowledge feature processing layer, the policy triggering behavior vector in the first original risk control policy response vector is updated using the second original risk control policy response vector targetvector_x to obtain a first optimized risk control policy response vector majorization vector_x corresponding to the risk control policy execution text to be analyzed, including:

[0018] In the group decision unit, a first vector difference value between the first response event sub-vector pending part_y and the second response event sub-vector target part_y is determined, and the first vector difference value is determined as an update coefficient corresponding to the second response event sub-vector target part_y;

[0019] Performing a global operation on the at least two second response event sub-vectors according to update coefficients corresponding to the at least two second response event sub-vectors, to obtain an interlock event representation vector interlock vector_y corresponding to the first response event sub-vector pending part_y;

[0020] In the optimization decision unit, a second vector difference value between the linkage event representation vector interlock vector_y and the first response event sub-vector pending part_y is determined;

[0021] performing knowledge optimization on the first response event sub-vector pending part_y according to the second vector difference value corresponding to the first response event sub-vector pending part_y to obtain a response event matching vector;

[0022] According to the response event matching vectors respectively corresponding to the at least two first response event sub-vectors, a first optimized risk control strategy response vector majorization vector_x associated with the target risk control strategy execution text set target set1_x is generated.

[0023] In some example embodiments, determining, using at least two optimized response vector binary pairs, a risk control strategy matching vector that reflects a commonality score between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed, includes:

[0024] Perform feature downsampling on the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and determine the second optimized risk control strategy response vector after feature downsampling as the sliding filter operator corresponding to the first optimized risk control strategy response vector majorization vector_x;

[0025] Perform sliding filtering on the first optimized risk control strategy response vector majorization vector_x according to the sliding filtering operator corresponding to the first optimized risk control strategy response vector majorization vector_x to obtain a commonality score between the target risk control strategy execution text set target set1_x and the risk control strategy execution text to be analyzed;

[0026] A risk control strategy matching vector is generated based on the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed.

[0027] In some example embodiments, performing response focusing processing on the first original risk control policy response vector to obtain a risk control policy response focusing vector corresponding to the risk control policy execution text to be analyzed includes:

[0028] Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a conditional knowledge feature processing module and a regression processing module;

[0029] Loading the first original risk control strategy response vector into the conditional knowledge feature processing module, and performing knowledge optimization on the first original risk control strategy response vector through the conditional knowledge feature processing module to obtain a proposed risk control strategy response vector corresponding to the risk control strategy execution text to be analyzed;

[0030] The proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed is loaded into the regression processing module. Through the regression processing module, the proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed is regressed to obtain the risk control policy response focus vector corresponding to the risk control policy execution text to be analyzed.

[0031] In some example embodiments, determining, based on the risk control policy matching vector and the risk control policy response focus vector, the risk control policy label data corresponding to the risk control policy execution text to be analyzed from at least two pieces of risk control policy label data includes:

[0032] Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a first decision tree module and a second decision tree module;

[0033] Loading the risk control strategy matching vector into the first decision tree module, and determining, in the first decision tree module, a first decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy matching vector;

[0034] Loading the risk control strategy response focus vector into the second decision tree module, and determining, in the second decision tree module, a second decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy response focus vector;

[0035] Adding the first decision matching feature and the second decision matching feature to obtain a label matching feature corresponding to the risk control policy execution text to be analyzed;

[0036] Based on the label matching feature, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed is determined from at least two risk control strategy label data.

[0037] In some example embodiments, obtaining a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed, and obtaining second original risk control policy response vectors corresponding to at least two target risk control policy execution text sets, respectively, include:

[0038] Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a vector mining module;

[0039] The risk control policy execution text to be analyzed and at least two target risk control policy execution text sets associated with the risk control policy execution text to be analyzed are both loaded into the vector mining module; the at least two target risk control policy execution text sets include a target risk control policy execution text set target set2_z, and the target risk control policy execution text set target set2_z includes W target risk control policy execution texts having the same risk control policy label data, where z is a positive integer and is not greater than the total number of target risk control policy execution text sets corresponding to the at least two target risk control policy execution text sets; and W is a positive integer;

[0040] Performing vector mining on the risk control policy execution text to be analyzed by the vector mining module to obtain a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed;

[0041] Perform vector mining on the target risk control strategy execution text set target set2_z to obtain the target response vectors to be processed corresponding to the W target risk control strategy execution texts;

[0042] Vector averaging is performed on the W target response vectors to be processed to obtain the second original risk control strategy response vector corresponding to the target risk control strategy execution text set target set2_z.

[0043] In the second aspect, an embodiment of the present invention also provides a risk control strategy analysis system, including a processing engine, a network module and a memory, wherein the processing engine and the memory communicate through the network module, and the processing engine is used to read a computer program from the memory and run it to implement the above method.

[0044] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above method when the computer program is executed.

[0045] Other features will be described in part in the following description. Those skilled in the art will discover these features in part upon inspection of the following description and drawings, or may learn these features through production or use. The features of the present invention may be realized and obtained by practicing or using the various aspects of the methods, tools, and combinations set forth in the detailed examples described below.

[0046] In an embodiment of the present invention, the risk control strategy analysis system can perform knowledge optimization on at least two basic response vector tuples respectively to improve the response detail output performance corresponding to the risk control strategy execution text to be analyzed and at least two target risk control strategy execution text sets, and obtain at least two optimized response vector tuples. Therefore, the risk control strategy matching vector generated by at least two optimized response vector tuples can accurately reflect the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed; at the same time, the risk control strategy analysis system can perform response focusing processing on the first original risk control strategy response vector to improve the response detail output performance of the risk control strategy execution text to be analyzed, and obtain a risk control strategy response focusing vector; in this way, the risk control strategy matching vector can accurately reflect the risk control strategy to be analyzed. The commonality scores corresponding to the execution text and at least two target risk control strategy execution text sets, and the risk control strategy response focusing vector can improve the response detail output performance of the risk control strategy execution text to be analyzed, and thus improve the matching analysis accuracy of the risk control strategy execution text to be analyzed. In this way, based on the risk control strategy matching vector and the risk control strategy response focusing vector, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed can be accurately determined in at least two risk control strategy label data, that is, the matching analysis accuracy of the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed can be improved, thereby realizing accurate and rapid label classification of the risk control strategy execution text to be analyzed, providing convenience for subsequent rapid and accurate calling and retrieval of the risk control strategy execution text to be analyzed, avoiding traversal of the entire text database, and reducing unnecessary resource overhead. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0048] The methods, systems, and / or programs in the accompanying drawings will be further described according to exemplary embodiments. These exemplary embodiments will be described in detail with reference to the drawings. These exemplary embodiments are non-limiting exemplary embodiments, wherein reference numerals represent similar structures in the various views of the drawings.

[0049] Figure 1 Schematic diagram of the hardware and software components of an exemplary risk control strategy analysis system according to some embodiments of the present invention.

[0050] Figure 2 1 is a flowchart of an exemplary AI-based risk control strategy analysis method and / or process according to some embodiments of the present invention. DETAILED DESCRIPTION

[0051] In order to better understand the above technical solution, the technical solution of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0052] In the following detailed description, numerous specific details are set forth by way of example in order to provide a thorough understanding of the relevant teachings. However, it will be apparent to one skilled in the art that the present invention can be practiced without these details. In other instances, well-known methods, procedures, systems, compositions, and / or circuits have been described at a relatively high level, without detail, to avoid unnecessarily obscuring aspects of the present invention.

[0053] These and other characteristics, the functions disclosed by the present invention, the methods of performing the functions of the related elements in the structure and the combination of parts and the economics of production may become more apparent in consideration of the following description with reference to the accompanying drawings, all of which form a part of the present invention. However, it is to be understood that the drawings are for illustration and description purposes only and are not intended to limit the scope of the present invention. It should be understood that these drawings are not drawn to scale. However, it should be clearly understood that the drawings are for illustration and description purposes only and are not intended to limit the scope of the present invention. It should be understood that these drawings are not drawn to scale.

[0054] The present invention uses flowcharts to illustrate the execution processes performed by the system according to embodiments of the present invention. It should be clearly understood that the execution processes of the flowcharts may not be executed in sequence. Rather, the execution processes may be executed in reverse order or simultaneously. In addition, at least one additional execution process may be added to the flowcharts. One or more execution processes may be deleted from the flowcharts.

[0055] Figure 1 This is a block diagram of the structural composition of a risk control strategy analysis system 100 according to some embodiments of the present invention. The risk control strategy analysis system 100 may include a processing engine 110, a network module 120 and a memory 130. The processing engine 110 and the memory 130 communicate through the network module 120.

[0056] The processing engine 110 may process relevant information and / or data to perform one or more functions described herein. For example, in some embodiments, the processing engine 110 may include at least one processing engine (e.g., a single-core processing engine or a multi-core processor). By way of example only, the processing engine 110 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a graphics processing unit (GPU), a physical processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a controller, a microcontroller unit, a reduced instruction set computer (RISC), a microprocessor, or the like, or any combination thereof.

[0057] The network module 120 can facilitate the exchange of information and / or data. In some embodiments, the network module 120 can be any type of wired or wireless network, or a combination thereof. By way of example only, the network module 120 can include a cable network, a wired network, a fiber optic network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a public telephone switched network (PSTN), a Bluetooth network, a wireless personal area network (WLAN), a near field communication (NFC) network, or any combination thereof. In some embodiments, the network module 120 can include at least one network access point. For example, the network module 120 can include a wired or wireless network access point, such as a base station and / or a network access point.

[0058] The memory 130 may be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. The memory 130 is used to store programs, and the processing engine 110 executes the programs after receiving an execution instruction.

[0059] I understand. Figure 1 The structure shown is for illustration only. The risk control strategy analysis system 100 may also include Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown. Figure 1 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0060] Figure 2 is a flowchart of an exemplary AI-based risk control strategy analysis method and / or process according to some embodiments of the present invention, wherein the AI-based risk control strategy analysis method is applied to Figure 1The risk control strategy analysis system 100 may further include the technical solutions described in steps 1 to 5.

[0061] Step 1: Obtain a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed, and obtain second original risk control policy response vectors corresponding to at least two target risk control policy execution text sets.

[0062] Among them, the risk control strategy label data corresponding to the at least two target risk control strategy execution text sets are different.

[0063] For example, the risk control policy execution text in embodiments of the present invention can be a descriptive text corresponding to a risk control policy designated for internet finance, e-commerce, digital office services, and the like. The risk control policy execution text can record the specific risk control actions to be taken under specific circumstances. Because the scenarios faced by risk control policies are relatively complex, it is essential to categorize different risk control policy execution texts. To this end, embodiments of the present invention can perform categorization (i.e., label matching) on ​​the risk control policy execution texts to be analyzed. Based on this, the risk control policy execution texts to be analyzed and the target set of risk control policy execution texts can both correspond to the same business scenario, such as an e-commerce business scenario.

[0064] Furthermore, the first raw risk control policy response vector can be understood as the raw risk control interaction features of the risk control policy execution text to be analyzed. Based on the interaction between request and response information, the risk control policy response vector can record the characteristics of different risk control policy execution processes, thereby highlighting the differences between different risk control policy execution processes. Based on this, the risk control policy label data can be understood as the category information of the risk control policy (for example, the risk control policy label data can include a fraud risk control label, a traffic attack risk control label, etc.).

[0065] Step 2: Perform knowledge optimization on the basic response vector tuples corresponding to at least two second original risk control strategy response vectors to obtain an optimized response vector tuple corresponding to each basic response vector tuple.

[0066] A basic response vector tuple includes the first original risk control strategy response vector and a second original risk control strategy response vector. In other words, the basic response vector tuple can be understood as an initial vector pair.

[0067] In the embodiment of the present invention, knowledge optimization can be understood as strengthening the risk control strategy response vector, which can improve the expressiveness of detailed features.

[0068] Step 3: Determine, by using at least two optimized response vector binary pairs, a risk control strategy matching vector that reflects the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed.

[0069] In the embodiment of the present invention, the commonality score can be understood as the commonality degree or similarity degree. Based on this, the risk control strategy matching vector is used to measure whether the risk control strategies are similar or matched, and can therefore be understood as a risk control strategy measurement vector.

[0070] Step 4: Perform response focusing processing on the first original risk control strategy response vector to obtain a risk control strategy response focusing vector corresponding to the risk control strategy execution text to be analyzed.

[0071] In an embodiment of the present invention, the response focus processing can be understood as an attention enhancement operation, which aims to focus and enhance the response behavior characteristics in the first original risk control strategy response vector, thereby obtaining the response enhancement characteristics of the risk control strategy (risk control strategy response focus vector).

[0072] Step 5: Determine the risk control policy label data corresponding to the risk control policy execution text to be analyzed from at least two risk control policy label data based on the risk control policy matching vector and the risk control policy response focus vector.

[0073] It can be understood that in an embodiment of the present invention, the risk control policy matching vector is determined based on the optimized response vector binary, and the risk control policy response focus vector is determined through response focus processing. Therefore, the risk control policy matching vector and the risk control policy response focus vector can improve the detailed characterization capability of the response behavior characteristics. In this way, when matching the risk control policy label data through the risk control policy matching vector and the risk control policy response focus vector, the risk control policy label data corresponding to the risk control policy execution text to be analyzed can be accurately and reliably determined, thereby realizing accurate label classification of the risk control policy execution text to be analyzed, providing convenience for subsequent rapid and accurate calling and retrieval of the risk control policy execution text to be analyzed, avoiding traversal of the entire text database, and reducing unnecessary resource overhead.

[0074] In some exemplary embodiments, the knowledge optimization of the basic response vector tuples corresponding to at least two second original risk control strategy response vectors described in step 2 to obtain optimized response vector tuples corresponding to each basic response vector tuple may include the technical solutions described in steps 21 to 24.

[0075] Step 21: Obtain a risk control strategy analysis algorithm.

[0076] Among them, the risk control strategy analysis algorithm can be a neural network model built and trained based on AI technology. The training and debugging process of the risk control strategy analysis algorithm can be implemented in combination with the idea of ​​loss function convergence, which will not be elaborated here. Furthermore, the risk control strategy analysis algorithm includes a knowledge feature processing module, which can be used for feature enhancement processing. In addition, the at least two target risk control strategy execution text sets include a target risk control strategy execution text set target set1_x, and the at least two second original risk control strategy response vectors include a second original risk control strategy response vector target vector_x corresponding to the target risk control strategy execution text set target set1_x, where x is a positive integer, and x is not greater than the total number corresponding to the at least two target risk control strategy execution text sets.

[0077] Step 22: Load the basic response vector double vector_x consisting of the second original risk control strategy response vector target vector_x and the first original risk control strategy response vector into the knowledge feature processing module.

[0078] Step 23: Through the knowledge feature processing module, the basic response vector binary doublevector_x is optimized to obtain the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and the first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed.

[0079] Among them, the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x is related to the risk control strategy execution text to be analyzed; the first optimized risk control strategy response vector majorization vector_x is related to the target risk control strategy execution text set target set1_x.

[0080] Step 24: The second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x and the first optimized risk control strategy response vector majorization vector_x are combined to form an optimized response vector tuple corresponding to the basic response vector tuple double vector_x.

[0081] It can be understood that by determining different basic response vector tuples and then combining the knowledge feature processing module for knowledge optimization, the optimized response vector tuple corresponding to the basic response vector tuple double vector_x can be determined completely and orderly, thereby providing a data basis for subsequent feature matching and label classification.

[0082] In some examples, the knowledge feature processing module (knowledge feature processing network) includes a first knowledge feature processing layer and a second knowledge feature processing layer. Based on this, the knowledge feature processing module, as described in step 23, performs knowledge optimization on the basic response vector double vector_x to obtain the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x and the first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed. This may include the technical solutions described in steps 231 and 232.

[0083] Step 231: In the first knowledge feature processing layer, the policy triggering behavior vector in the first original risk control policy response vector is updated through the second original risk control policy response vector target vector_x to obtain the first optimized risk control policy response vector majorization vector_x corresponding to the risk control policy execution text to be analyzed.

[0084] Among them, the policy triggering behavior vector in the first original risk control policy response vector can be strengthened based on the second original risk control policy response vector target vector_x. The policy triggering behavior vector focuses on the before and after timing characteristics of the business request when triggering the risk control policy, so it can accurately and completely realize the feature enhancement at the risk control policy response level.

[0085] Step 232: In the second knowledge feature processing layer, the policy triggering behavior vector in the second original risk control policy response vector target vector_x is updated using the first original risk control policy response vector to obtain the second optimized risk control policy response vector corresponding to the target risk control policy execution text set target set1_x.

[0086] It can be understood that based on step 231 and step 232, feature enhancement processing can be achieved between each other, thereby accurately and completely achieving feature enhancement at the risk control strategy response level.

[0087] In some other exemplary embodiments, the first knowledge feature processing layer includes a clustering decision unit (clustering unit) and an optimization decision unit (vector difference operation unit). Furthermore, the second original risk control strategy response vector target vector_x includes at least two second response event subvectors; the at least two second response event subvectors include a second response event subvector target part_y; the first original risk control strategy response vector includes at least two first response event subvectors; the at least two first response event subvectors include a first response event subvector pending part_y, where y is a positive integer and not greater than the total number of corresponding first response event subvectors. Based on this, step 231, as described in the first knowledge feature processing layer, updates the policy triggering behavior vector in the first original risk control strategy response vector using the second original risk control strategy response vector target vector_x to obtain the first optimized risk control strategy response vector majorizationvector_x corresponding to the risk control strategy execution text to be analyzed. This may include the technical solutions described in steps 2311-2315.

[0088] Step 2311: In the group decision unit, determine the first vector difference value between the first response event sub-vector pending part_y and the second response event sub-vector target part_y, and determine the first vector difference value as the update coefficient corresponding to the second response event sub-vector target part_y.

[0089] The vector difference value can be understood as the feature difference or feature distance, and the update coefficient can be understood as the enhancement coefficient or enhancement weight.

[0090] Step 2312: Perform a global operation (which can be understood as a weighted summation process) on the at least two second response event sub-vectors based on the update coefficients corresponding to the at least two second response event sub-vectors, and obtain the linkage event representation vector interlock vector_y corresponding to the first response event sub-vector pending part_y.

[0091] The linkage event representation vector interlock vector_y can be understood as a weighted vector or a vector with similar characteristics.

[0092] Step 2313: In the optimization decision unit, determine a second vector difference value between the linkage event representation vector interlockvector_y and the first response event sub-vector pending part_y.

[0093] Step 2314: perform knowledge optimization on the first response event sub-vector pending part_y according to the second vector difference value corresponding to the first response event sub-vector pending part_y to obtain a response event matching vector.

[0094] Among them, the response event matching vector focuses on the detailed representation of the response results of the risk control strategy. For example, the characteristic information corresponding to the permission authentication event triggered by a certain access behavior can be recorded through the response event matching vector.

[0095] Step 2315: Generate a first optimized risk control strategy response vector majorization vector_x associated with the target risk control strategy execution text set target set1_x based on the response event matching vectors corresponding to the at least two first response event sub-vectors.

[0096] It can be understood that based on steps 2311 to 2315, the cluster decision unit and the optimization decision unit can be used to perform corresponding feature difference operations, feature fusion, and feature enhancement, so as to quickly and accurately obtain the first optimized risk control strategy response vector majorization vector_x that is related to the target risk control strategy execution text set target set1_x.

[0097] Under some possible design ideas, the method described in step 3 of determining the risk control strategy matching vector used to reflect the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed through at least two optimized response vector tuples may include the technical solutions described in steps 31 to 33.

[0098] Step 31: Perform feature downsampling on the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and determine the second optimized risk control strategy response vector after feature downsampling as the sliding filter operator corresponding to the first optimized risk control strategy response vector majorization vector_x.

[0099] Among them, feature downsampling can be understood as pooling processing. After feature downsampling, the convolution operator (sliding filter operator / convolution kernel) determined by the corresponding second optimized risk control strategy response vector can be used.

[0100] Step 32: Perform sliding filtering on the first optimized risk control strategy response vector majorization vector_x according to the sliding filtering operator corresponding to the first optimized risk control strategy response vector majorization vector_x to obtain a commonality score between the target risk control strategy execution text set target set1_x and the risk control strategy execution text to be analyzed.

[0101] It can be understood that the sliding filter operator can achieve targeted convolution processing, thereby reducing the interference of redundant features and noise features to ensure the accuracy of the determined commonality score.

[0102] Step 33: Generate a risk control strategy matching vector based on the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed.

[0103] Applied to steps 31 to 33, after feature downsampling, the convolution operator (sliding filter operator / convolution kernel) determined by the corresponding second optimized risk control strategy response vector can be used, and targeted convolution processing can be achieved through the sliding filter operator, thereby reducing the interference of redundant features and noise features to ensure the accuracy of the determined commonality score. In this way, when generating the risk control strategy matching vector, the accuracy and credibility of the risk control strategy matching vector can be ensured.

[0104] Under some possible design ideas, the response focusing processing of the first original risk control strategy response vector described in step 4 to obtain the risk control strategy response focusing vector corresponding to the risk control strategy execution text to be analyzed may include the technical solutions described in steps 41 to 43.

[0105] Step 41: Obtain a risk control strategy analysis algorithm.

[0106] Among them, the risk control strategy analysis algorithm can also include a conditional knowledge feature processing module (autocorrelation knowledge feature processing module) and a regression processing module (linear processing module).

[0107] Step 42: Load the first original risk control strategy response vector into the conditional knowledge feature processing module. Through the conditional knowledge feature processing module, perform knowledge optimization on the first original risk control strategy response vector to obtain the proposed processed risk control strategy response vector corresponding to the risk control strategy execution text to be analyzed.

[0108] Based on this step, the feature enhancement processing of autocorrelation can be achieved, thereby avoiding some disturbances caused by cross-correlation enhancement.

[0109] Step 43: Load the proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed into the regression processing module. Through the regression processing module, the proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed is regressed to obtain the risk control policy response focus vector corresponding to the risk control policy execution text to be analyzed.

[0110] Based on this step, linear processing of the risk control strategy response vector to be processed can be achieved, thereby improving the feature expression capability of the risk control strategy response focus vector.

[0111] It can be understood that the application of steps 41 to 43 can combine the conditional knowledge feature processing module (autocorrelation knowledge feature processing module) and the regression processing module (linear processing module) to achieve autocorrelation feature enhancement processing to avoid some disturbances caused by cross-correlation enhancement, and achieve linear processing of the risk control strategy response vector to be processed to improve the feature expression ability of the risk control strategy response focus vector, so as to ensure the feature detail quality of the generated risk control strategy response focus vector.

[0112] Under other possible design ideas, the method described in step 5, based on the risk control strategy matching vector and the risk control strategy response focus vector, determines the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed in at least two risk control strategy label data, and may include the technical solutions described in steps 51 to 55.

[0113] Step 51: Obtain a risk control strategy analysis algorithm.

[0114] The risk control strategy analysis algorithm includes a first decision tree module and a second decision tree module. For example, the first decision tree module can be a metric decision module, and the second decision tree module can be a fine-tuning decision module. Furthermore, the decision tree module can also be understood as a classifier or classification module.

[0115] Step 52: Load the risk control strategy matching vector into the first decision tree module. In the first decision tree module, determine a first decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy matching vector.

[0116] In the embodiment of the present invention, the first decision matching feature can be understood as a first classification prediction vector.

[0117] Step 53: Load the risk control strategy response focus vector into the second decision tree module. In the second decision tree module, determine the second decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy response focus vector.

[0118] In the embodiment of the present invention, the second decision matching feature can be understood as a second classification prediction vector.

[0119] Step 54: Add the first decision matching feature and the second decision matching feature to obtain a label matching feature corresponding to the risk control policy execution text to be analyzed.

[0120] By summing the first decision matching feature and the second decision matching feature, a label matching feature (which can also be understood as a global category prediction vector) that is as accurate and complete as possible can be obtained.

[0121] Step 55: Determine, based on the label matching feature, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed from at least two risk control strategy label data.

[0122] It can be understood that decision matching features are determined through different types of decision tree modules, and feature summation can be performed based on these decision matching features to obtain label matching features that are as accurate and complete as possible. Therefore, based on the label matching features that are as accurate and complete as possible, accurate label matching and classification of the risk control strategy execution text to be analyzed can be achieved.

[0123] Under some exemplary design ideas, the step 1 described in obtaining the first original risk control strategy response vector corresponding to the risk control strategy execution text to be analyzed, and obtaining the second original risk control strategy response vector corresponding to at least two target risk control strategy execution text sets respectively, may include the technical solutions described in steps 11 to 15.

[0124] Step 11: Obtain the risk control strategy analysis algorithm.

[0125] Among them, the risk control strategy analysis algorithm includes a vector mining module (feature extraction module).

[0126] Step 12: Load the risk control policy execution text to be analyzed and at least two target risk control policy execution text sets that are related to the risk control policy execution text to be analyzed into the vector mining module.

[0127] Among them, the at least two target risk control strategy execution text sets include a target risk control strategy execution text set target set2_z, and the target risk control strategy execution text set target set2_z includes W target risk control strategy execution texts with the same risk control strategy label data, z is a positive integer, and z is not greater than the total number corresponding to the at least two target risk control strategy execution text sets; W is a positive integer.

[0128] Step 13: Perform vector mining on the risk control policy execution text to be analyzed by the vector mining module to obtain a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed.

[0129] Step 14: Perform vector mining on the target risk control strategy execution text set target set2_z to obtain the target response vectors to be processed corresponding to the W target risk control strategy execution texts.

[0130] Step 15: Perform vector averaging on the W target response vectors to be processed to obtain the second original risk control strategy response vector corresponding to the target risk control strategy execution text set target set2_z.

[0131] It can be understood that different mining processing methods are used to obtain the first original risk control strategy response vector and the second original risk control strategy response vector, which can take into account the differences between a single vector and a vector group, thereby ensuring the accuracy and completeness of the first original risk control strategy response vector and the second original risk control strategy response vector.

[0132] In an embodiment of the present invention, the risk control strategy analysis system can perform knowledge optimization on at least two basic response vector tuples respectively to improve the response detail output performance corresponding to the risk control strategy execution text to be analyzed and at least two target risk control strategy execution text sets, and obtain at least two optimized response vector tuples. Therefore, the risk control strategy matching vector generated by at least two optimized response vector tuples can accurately reflect the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed; at the same time, the risk control strategy analysis system can perform response focusing processing on the first original risk control strategy response vector to improve the response detail output performance of the risk control strategy execution text to be analyzed, and obtain a risk control strategy response focusing vector; in this way, the risk control strategy matching vector can accurately reflect the risk control strategy to be analyzed. The commonality scores corresponding to the execution text and at least two target risk control strategy execution text sets, and the risk control strategy response focusing vector can improve the response detail output performance of the risk control strategy execution text to be analyzed, and thus improve the matching analysis accuracy of the risk control strategy execution text to be analyzed. In this way, based on the risk control strategy matching vector and the risk control strategy response focusing vector, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed can be accurately determined in at least two risk control strategy label data, that is, the matching analysis accuracy of the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed can be improved, thereby realizing accurate and rapid label classification of the risk control strategy execution text to be analyzed, providing convenience for subsequent rapid and accurate calling and retrieval of the risk control strategy execution text to be analyzed, avoiding traversal of the entire text database, and reducing unnecessary resource overhead.

[0133] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed description provided above is merely illustrative and does not limit the present invention. Although not explicitly described herein, various modifications, improvements, and revisions to the present invention may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in the present invention and remain within the spirit and scope of the exemplary embodiments of the present invention.

[0134] At the same time, the present invention uses specific terms to describe embodiments of the present invention. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a certain feature, structure, or characteristic associated with at least one embodiment of the present invention. Therefore, it should be emphasized and noted that the mention of "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different parts of this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics of at least one embodiment of the present invention may be appropriately combined.

[0135] Furthermore, those skilled in the art will appreciate that various aspects of the present invention may be illustrated and described in terms of a number of patentable categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvement thereof. Accordingly, various aspects of the present invention may be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. Each of the above hardware and software may be referred to as a "unit," "component," or "system." Furthermore, various aspects of the present invention may be embodied as a computer product embodied in at least one computer-readable medium, the product comprising computer-readable program code.

[0136] A computer-readable storage medium may comprise a propagated data signal embodying computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may be in a variety of forms, including electromagnetic, optical, etc., or a suitable combination thereof. A computer-readable storage medium may be any computer-readable medium other than a computer-readable storage medium that can be connected to an instruction execution system, apparatus, or device to communicate, propagate, or transmit the program for use. The program code on the computer-readable storage medium may be transmitted via any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above.

[0137] The computer program code required to implement various aspects of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, or similar conventional programming languages ​​such as the C programming language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby, and Groovy, or other programming languages. The programming code may be executed entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer via any network, such as a local area network (LAN) or a wide area network (WAN), or to an external computer (e.g., via the Internet), or in a cloud computing environment, or as a service such as software as a service (SaaS). Furthermore, unless expressly provided in the claims, the order of the processing elements and sequences, the use of numerals, or other designations described herein are not intended to limit the order of the processes and methods of the present invention. Although the foregoing disclosure discusses some currently believed useful embodiments of the invention through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. Rather, the claims are intended to cover all modifications and equivalent combinations consistent with the spirit and scope of the embodiments of the invention. For example, while the system components described above can be implemented using hardware devices, they can also be implemented using software-only solutions, such as installing the described system on an existing server or mobile device. It should also be understood that in order to simplify the presentation of the present disclosure and to facilitate understanding of at least one embodiment of the invention, the foregoing description of the embodiments of the invention sometimes combines multiple features into a single embodiment, figure, or description thereof. However, this method of disclosure does not imply that the subject matter of the invention requires more features than those recited in the claims. In fact, the features of an embodiment may be fewer than all the features of the individual embodiments disclosed above.

Claims

1. An AI-based risk control strategy analysis method, characterized in that: Applied to a risk control strategy analysis system, the method includes: Obtaining a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed, and obtaining second original risk control policy response vectors corresponding to at least two target risk control policy execution text sets, respectively; wherein the risk control policy label data corresponding to the at least two target risk control policy execution text sets are different; Obtain a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a knowledge feature processing module; the at least two target risk control strategy execution text sets include a target risk control strategy execution text set target set1_x, and at least two second original risk control strategy response vectors include a second original risk control strategy response vector target vector_x corresponding to the target risk control strategy execution text set target set1_x, where x is a positive integer and is not greater than the total number of corresponding at least two target risk control strategy execution text sets; a basic response vector tuple double vector_x consisting of the second original risk control strategy response vector target vector_x and the first original risk control strategy response vector is loaded into the knowledge feature processing module; through the knowledge feature processing module, the basic response vector tuple doublevector_x is knowledge optimized to obtain a second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and a first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed; wherein, the target risk control strategy execution text set target The second optimized risk control strategy response vector corresponding to set1_x is associated with the risk control strategy execution text to be analyzed; the first optimized risk control strategy response vector majorization vector_x is associated with the target risk control strategy execution text set targetset1_x; the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set targetset1_x and the first optimized risk control strategy response vector majorization vector_x are combined to form an optimized response vector tuple corresponding to the basic response vector tuple double vector_x; wherein, a basic response vector tuple includes the first original risk control strategy response vector and a second original risk control strategy response vector; Perform feature downsampling on the second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x, and determine the second optimized risk control strategy response vector after feature downsampling as the sliding filter operator corresponding to the first optimized risk control strategy response vector majorization vector_x; perform sliding filtering processing on the first optimized risk control strategy response vector majorization vector_x according to the sliding filter operator corresponding to the first optimized risk control strategy response vector majorization vector_x to obtain the commonality score between the target risk control strategy execution text set target set1_x and the risk control strategy execution text to be analyzed; generate a risk control strategy matching vector based on the commonality scores between the at least two target risk control strategy execution text sets and the risk control strategy execution text to be analyzed; perform response focusing processing on the first original risk control strategy response vector to obtain a risk control strategy response focusing vector corresponding to the risk control strategy execution text to be analyzed; According to the risk control policy matching vector and the risk control policy response focus vector, the risk control policy label data corresponding to the risk control policy execution text to be analyzed is determined from at least two risk control policy label data.

2. The method according to claim 1, characterized in that The knowledge feature processing module includes a first knowledge feature processing layer and a second knowledge feature processing layer. The knowledge feature processing module performs knowledge optimization on the basic response vector double vector_x to obtain a second optimized risk control strategy response vector corresponding to the target risk control strategy execution text set target set1_x and a first optimized risk control strategy response vector majorization vector_x corresponding to the risk control strategy execution text to be analyzed, including: In the first knowledge feature processing layer, the policy triggering behavior vector in the first original risk control policy response vector is updated using the second original risk control policy response vector target vector_x to obtain a first optimized risk control policy response vector majorization vector_x corresponding to the risk control policy execution text to be analyzed; In the second knowledge feature processing layer, the policy triggering behavior vector in the second original risk control policy response vector target vector_x is updated through the first original risk control policy response vector to obtain the second optimized risk control policy response vector corresponding to the target risk control policy execution text set target set1_x.

3. The method according to claim 2, characterized in that The first knowledge feature processing layer includes a group decision unit and an optimization decision unit; the second original risk control strategy response vector target vector_x includes at least two second response event sub-vectors; the at least two second response event sub-vectors include a second response event sub-vector targetpart_y; the first original risk control strategy response vector includes at least two first response event sub-vectors; the at least two first response event sub-vectors include a first response event sub-vector pending part_y, where y is a positive integer and is not greater than the total number corresponding to the at least two first response event sub-vectors; In the first knowledge feature processing layer, the policy triggering behavior vector in the first original risk control policy response vector is updated using the second original risk control policy response vector targetvector_x to obtain a first optimized risk control policy response vector majorization vector_x corresponding to the risk control policy execution text to be analyzed, including: In the group decision unit, a first vector difference value between the first response event sub-vector pending part_y and the second response event sub-vector target part_y is determined, and the first vector difference value is determined as an update coefficient corresponding to the second response event sub-vector target part_y; Performing a global operation on the at least two second response event sub-vectors according to update coefficients corresponding to the at least two second response event sub-vectors, to obtain an interlock event representation vector interlock vector_y corresponding to the first response event sub-vector pending part_y; In the optimization decision unit, a second vector difference value between the linkage event representation vector interlock vector_y and the first response event sub-vector pending part_y is determined; performing knowledge optimization on the first response event sub-vector pending part_y according to the second vector difference value corresponding to the first response event sub-vector pending part_y to obtain a response event matching vector; According to the response event matching vectors respectively corresponding to the at least two first response event sub-vectors, a first optimized risk control strategy response vector majorization vector_x associated with the target risk control strategy execution text set target set1_x is generated.

4. The method according to claim 1, wherein The performing response focusing processing on the first original risk control strategy response vector to obtain a risk control strategy response focusing vector corresponding to the risk control strategy execution text to be analyzed includes: Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a conditional knowledge feature processing module and a regression processing module; Loading the first original risk control strategy response vector into the conditional knowledge feature processing module, and performing knowledge optimization on the first original risk control strategy response vector through the conditional knowledge feature processing module to obtain a proposed risk control strategy response vector corresponding to the risk control strategy execution text to be analyzed; The proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed is loaded into the regression processing module. Through the regression processing module, the proposed risk control policy response vector corresponding to the risk control policy execution text to be analyzed is regressed to obtain the risk control policy response focus vector corresponding to the risk control policy execution text to be analyzed.

5. The method according to claim 1, wherein The step of determining, based on the risk control strategy matching vector and the risk control strategy response focus vector, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed from at least two pieces of risk control strategy label data includes: Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a first decision tree module and a second decision tree module; Loading the risk control strategy matching vector into the first decision tree module, and determining, in the first decision tree module, a first decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy matching vector; Loading the risk control strategy response focus vector into the second decision tree module, and determining, in the second decision tree module, a second decision matching feature corresponding to the risk control strategy execution text to be analyzed based on the risk control strategy response focus vector; Adding the first decision matching feature and the second decision matching feature to obtain a label matching feature corresponding to the risk control policy execution text to be analyzed; Based on the label matching feature, the risk control strategy label data corresponding to the risk control strategy execution text to be analyzed is determined from at least two risk control strategy label data.

6. The method according to claim 1, characterized in that The step of obtaining a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed and obtaining second original risk control policy response vectors corresponding to at least two target risk control policy execution text sets respectively includes: Obtaining a risk control strategy analysis algorithm; the risk control strategy analysis algorithm includes a vector mining module; The risk control policy execution text to be analyzed and at least two target risk control policy execution text sets associated with the risk control policy execution text to be analyzed are both loaded into the vector mining module; the at least two target risk control policy execution text sets include a target risk control policy execution text set target set2_z, and the target risk control policy execution text set target set2_z includes W target risk control policy execution texts having the same risk control policy label data, where z is a positive integer and is not greater than the total number of target risk control policy execution text sets corresponding to the at least two target risk control policy execution text sets; and W is a positive integer; Performing vector mining on the risk control policy execution text to be analyzed by the vector mining module to obtain a first original risk control policy response vector corresponding to the risk control policy execution text to be analyzed; Perform vector mining on the target risk control strategy execution text set target set2_z to obtain the target response vectors to be processed corresponding to the W target risk control strategy execution texts; Vector averaging is performed on the W target response vectors to be processed to obtain the second original risk control strategy response vector corresponding to the target risk control strategy execution text set target set2_z.

7. A risk control strategy analysis system, characterized in that: The method comprises a processing engine, a network module and a memory, wherein the processing engine and the memory communicate with each other through the network module, and the processing engine is used to read a computer program from the memory and run it to implement the method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is run, the method according to any one of claims 1 to 6 is implemented.

Images