CN117094555A

CN117094555A - Updating method, device and equipment of wind control strategy

Info

Publication number: CN117094555A
Application number: CN202311007505.6A
Authority: CN
Inventors: 王立; 陆毅成
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2023-08-10
Filing date: 2023-08-10
Publication date: 2023-11-21

Abstract

The embodiment of the specification discloses a method, a device and equipment for updating a wind control strategy, wherein the method comprises the following steps: acquiring historical service data containing a first feature, and performing equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain values of a plurality of different cutting points corresponding to the first feature; based on the numerical value of each cutting point, constructing two mutually exclusive strategy conditions, and acquiring the risk leakage rate and/or the risk management rate corresponding to each strategy condition; acquiring target strategy conditions of which the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate; based on the target strategy conditions, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions to the service wind control strategies corresponding to the historical service data or not based on the evaluation result.

Description

Updating method, device and equipment of wind control strategy

Technical Field

The present document relates to the field of computer technologies, and in particular, to a method, an apparatus, and a device for updating a wind control policy.

Background

Generally, the wind control strategy mainly relies on manual analysis of related data, and then some combinations of strategy rules are generated through manual experience according to the black-producing technique and the like. In general, when new risk events are faced, and the situation that people pay more attention to private data is considered, the corresponding countermeasures can be made according to the manual experience to quickly respond, so that the risk is reduced in time. However, the manner of artificial experience is easily limited by personal cognition and experience, and is easily influenced by human prejudice and mistakes, thereby causing uncertainty and misjudgment and limiting the accuracy and reliability of the wind control strategy. Therefore, a better strategy upgrading iteration scheme is needed to be provided, so that the accuracy of the strategy can be improved, the user experience can be optimized, and unnecessary resource expenses such as follow-up risk management and control, user complaints, manual limitation solving and the like caused by strategy misaudit can be reduced.

Disclosure of Invention

The embodiment of the specification aims to provide a more optimal strategy upgrading iteration scheme, which not only can improve the accuracy of the strategy and optimize the user experience, but also can reduce unnecessary resource expenses such as follow-up risk management and control, user complaints, manual limitation solving and the like caused by strategy misaudit.

In order to achieve the above technical solution, the embodiments of the present specification are implemented as follows:

the embodiment of the specification provides a method for updating a wind control strategy, which comprises the following steps: and acquiring historical service data containing a first feature, and performing equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature. Based on the numerical value of each cutting point, two mutually exclusive strategy conditions are constructed, and the risk leakage rate and/or the risk management rate corresponding to each strategy condition are obtained. And acquiring target strategy conditions of which the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate. Based on the target strategy conditions, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions to the service wind control strategies corresponding to the historical service data or not based on the evaluation result.

The embodiment of the present disclosure provides an apparatus for updating a wind control policy, where the apparatus includes: and the characteristic cutting module is used for obtaining historical service data containing a first characteristic, and carrying out equal-frequency cutting on the range of the characteristic value corresponding to the first characteristic in the historical service data to obtain the numerical values of a plurality of different cutting points corresponding to the first characteristic. And the strategy construction module is used for constructing two mutually exclusive strategy conditions based on the numerical value of each cutting point and acquiring the risk leakage rate and/or the risk management rate corresponding to each strategy condition. And the screening module is used for acquiring target strategy conditions of which the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate. And the wind control strategy updating module is used for determining an evaluation result corresponding to the target strategy condition through a pre-trained strategy evaluation model based on the target strategy condition, and determining whether to add the target strategy condition into the service wind control strategy corresponding to the historical service data based on the evaluation result.

The embodiment of the present disclosure provides an update device for a wind control policy, where the update device for a wind control policy includes: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to: and acquiring historical service data containing a first feature, and performing equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature. Based on the numerical value of each cutting point, two mutually exclusive strategy conditions are constructed, and the risk leakage rate and/or the risk management rate corresponding to each strategy condition are obtained. And acquiring target strategy conditions of which the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate. Based on the target strategy conditions, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions to the service wind control strategies corresponding to the historical service data or not based on the evaluation result.

The present description also provides a storage medium for storing computer-executable instructions that when executed by a processor implement the following: and acquiring historical service data containing a first feature, and performing equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature. Based on the numerical value of each cutting point, two mutually exclusive strategy conditions are constructed, and the risk leakage rate and/or the risk management rate corresponding to each strategy condition are obtained. And acquiring target strategy conditions of which the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate. Based on the target strategy conditions, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions to the service wind control strategies corresponding to the historical service data or not based on the evaluation result.

Drawings

For a clearer description of embodiments of the present description or of the solutions of the prior art, the drawings that are required to be used in the description of the embodiments or of the prior art will be briefly described, it being obvious that the drawings in the description below are only some of the embodiments described in the description, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art;

FIG. 1 is a diagram illustrating an embodiment of a method for updating a wind control strategy according to the present disclosure;

FIG. 2 is a schematic diagram of another embodiment of a method for updating a pneumatic control strategy according to the present disclosure;

FIG. 3 is a schematic diagram of a model training process of a strategy evaluation model according to the present disclosure;

FIG. 4 is a schematic diagram of another embodiment of a method for updating a pneumatic control strategy according to the present disclosure;

FIG. 5 is a schematic diagram of an update process of the wind control strategy according to the present disclosure;

FIG. 6 is an embodiment of an update apparatus for a pneumatic control strategy according to the present disclosure;

fig. 7 is an embodiment of an update apparatus for a wind control strategy according to the present disclosure.

Detailed Description

The embodiment of the specification provides a method, a device and equipment for updating a wind control strategy.

In order to make the technical solutions in the present specification better understood by those skilled in the art, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present specification, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are intended to be within the scope of the present disclosure.

The embodiment of the specification provides a strategy upgrading iteration scheme, and generally, a wind control strategy mainly depends on manual analysis of related data, and then a combination of strategy rules is produced through manual experience according to a black-yield method and the like. In general, when a new risk event is faced, a response can be quickly made according to manual experience, and corresponding countermeasures are made, so that the risk is timely reduced. However, the manner of artificial experience is easily limited by personal cognition and experience, and is easily influenced by human prejudice and mistakes, thereby causing uncertainty and misjudgment and limiting the accuracy and reliability of the wind control strategy. Therefore, the scheme carries out upgrading iteration on the stock strategy through intelligence, on one hand, the accuracy of the strategy can be improved through a large amount of data support, and the user experience is optimized; on the other hand, unnecessary resource expenses such as follow-up risk management and control, user complaints, manual limitation solving and the like caused by strategy misaudit are reduced. Specific processing can be seen from the details in the following examples.

Example 1

As shown in fig. 1, the embodiment of the present disclosure provides a method for updating a wind control policy, where an execution subject of the method may be a terminal device or a server, where the terminal device may be a mobile terminal device such as a mobile phone, a tablet computer, or a computer device such as a notebook computer or a desktop computer, or may also be an IoT device (specifically, such as a smart watch, a vehicle-mounted device, or the like), and where the server may be a separate server, or may be a server cluster formed by a plurality of servers, and the server may be a background server such as a financial service or an online shopping service, or may be a background server of a certain application program, or the like. In this embodiment, the execution subject is taken as a server for example for detailed description, and for the case that the execution subject is a terminal device, the following processing of the case of the server may be referred to, and will not be described herein. The method specifically comprises the following steps:

In step S102, historical service data including the first feature is obtained, and the range of feature values corresponding to the first feature in the historical service data is subjected to equal-frequency cutting to obtain the numerical values of a plurality of different cutting points corresponding to the first feature.

The first feature may be any feature in the features included in the historical service data, where the features in this embodiment may include two parts, one part is information such as an identifier, and the other part is a feature value, for example, for the first feature, the first feature may include information such as an identifier of the first feature, where the first feature may represent the first feature, and specifically, the first feature may be an average historical amount, and the feature value corresponding to the first feature may be 230 yuan. The historical service data may be historical data related to a specified target service, for example, the target service may be a payment service, the historical service data may include historical service data of the payment service, and specifically may include, for example, payment account information, collection account information, payment time, payment location, payment amount, and collection account historical risk information, which may be specifically set according to practical situations, and the embodiment of the present disclosure is not limited to this.

In implementation, for a specified target service, the server may record relevant data (i.e., service data) generated during the execution of the target service by the user each time the user executes the target service, where the relevant data may include sequence data of operation actions of the user, relevant data of the target service, relevant data input by the user, relevant data selected by the user, relevant data provided by the server of the target service, and other service data. After a certain period of time (such as a month before the current time, a 6 month before the current time or a year before the current time) of data accumulation, a certain amount of historical service data with a certain data volume can be recorded in the server. When it is required to determine whether the feature included in the history service data is a newly added feature, whether the corresponding wind control policy needs to update the current risk policy, if so, the history service data or a part of the history service data therein with a certain amount and a certain data amount may be obtained, then the history service data or a part of the history service data therein with a certain amount and a certain data amount may be analyzed, the feature included in the history service data may be obtained through the analysis result, a range of feature values corresponding to each feature may also be determined, for example, the feature of the history average amount may be included in the history service data, the feature values may include 105,110, 102, 122, 154, … 135, the minimum value of the feature values is 102, and the maximum value of the feature values is 154, the range of feature values may be {102, … 105, … 110, …, …,135, … } or [102,154], and the range of feature values corresponding to the first feature in the historical service data may be subjected to equal-frequency cutting according to practical situations, for example, the range of feature values corresponding to the first feature may be subjected to equal-frequency cutting into 5 parts, specifically, for example {102, … }, {105, … }, {110, … }, {122, … }, {135, … } or [102,105], (105, 110], (110, 122], (122,135), and (135,154) }, and the values of the nodes including the values of 105,110, 122,135, 154 and the like may be obtained based on the above examples.

In step S104, two mutually exclusive policy conditions are constructed based on the value of each cutting point, and the risk leakage rate and/or the risk management rate corresponding to each policy condition are obtained.

The risk leakage rate may be used to represent a probability that the preset risk is not detected by the policy condition, and the risk management rate may be used to represent a probability that the user is not affected in using the corresponding business service, where in practical application, the smaller the risk leakage rate, the better the risk management rate is, and the smaller the risk management rate is.

In implementation, when the value of a certain cutting point is greater than or equal to the value, a certain condition may be satisfied, and when the value may be less than the value, a certain condition may be satisfied, so that two mutually exclusive policy conditions may be constructed, specifically, for example, the condition is a free riding train, and if the value of a certain cutting point is 1.5 meters, two mutually exclusive policy conditions such as a free riding train when the height is greater than or equal to 1.5 meters, and a free riding train when the height is less than 1.5 meters may be constructed. The foregoing is only one possible processing manner, and in practical application, a plurality of different processing manners may be further included, specifically, for example, a corresponding result may be calculated by a preset algorithm based on a value of a certain cutting point, two mutually exclusive policy conditions may be constructed by the result, for example, based on the foregoing example, a value of a certain cutting point is 1.5 meters, a corresponding calculation may be performed on 1.5 by a preset algorithm, and the obtained result may be 1.2 meters, so that two mutually exclusive policy conditions such as a free train may be constructed when the height is greater than or equal to 1.2 meters, and a free train may be constructed when the height is less than 1.2 meters, or other various realizable manners may be further included, and may be specifically set according to practical situations, where embodiments of the present specification do not limit this.

By the method, two mutually exclusive strategy conditions can be constructed based on the numerical value of each cutting point, so that multiple strategy conditions can be obtained, each strategy condition can be set in the target service for determining the matching degree of each strategy condition and the target service, the strategy condition can be tested in the process of using the target service in real time by a user, after a certain period of time is tested, the risk leakage rate and/or the risk management rate corresponding to each strategy condition can be counted, and therefore the risk leakage rate and/or the risk management rate corresponding to each strategy condition can be obtained, or the designated test data can be used for carrying out simulation test on the strategy condition, and the risk leakage rate and/or the risk management rate corresponding to each strategy condition can be counted, so that the risk leakage rate and/or the risk management rate corresponding to each strategy condition can be obtained.

In step S106, a target policy condition that the risk leakage rate and/or the risk management rate satisfy a preset policy screening condition is obtained from the constructed policy conditions, where the policy screening condition is constructed based on the risk leakage rate and/or the risk management rate.

In implementation, policy screening conditions constructed by a threshold value of risk leakage rate and/or a threshold value of risk management rate and the like may be preset according to actual situations, specifically, for example, when the risk leakage rate is smaller than a first threshold value and the risk management rate is smaller than a second threshold value, the policy conditions are selected, otherwise, the policy conditions are refused to be selected, and the policy screening conditions may include one or a plurality of policy screening conditions, and may be specifically set according to actual situations, which is not limited in the embodiment of the present specification. After a plurality of different strategy conditions are constructed in the mode, the strategy screening conditions can be used for screening the strategy conditions, so that the target strategy conditions that the risk leakage rate and/or the risk management rate obtained from the constructed strategy conditions meet the preset strategy screening conditions are obtained.

In step S108, based on the target policy condition, an evaluation result corresponding to the target policy condition is determined by a pre-trained policy evaluation model, and based on the evaluation result, it is determined whether to add the target policy condition to the service wind control policy corresponding to the historical service data.

The policy evaluation model may be a model for evaluating the constructed policy conditions, and the policy evaluation model may be constructed by various algorithms or models, for example, may be constructed by a BERT model or the like similar to an MLM model, or may also be constructed by a neural network algorithm or the like, and may be specifically set according to actual situations.

In implementation, a corresponding algorithm or model may be obtained, and a policy evaluation model may be constructed based on the algorithm or model, where input data of the policy evaluation model may be content of a policy condition, and output data may be an evaluation result of evaluating the policy condition, for example, the policy evaluation model may include a BERT model, input data of the policy evaluation model may be content of the policy condition, and finally the BERT model calculates and outputs an evaluation result of evaluating the policy condition and so on. Then, a training sample (i.e., a training sample constructed by the policy conditions, etc.) for training the policy evaluation model may be obtained, the model training may be performed using the training sample to perform model training, an objective function may be preset during the model training, and optimization processing may be performed based on model parameters in the policy evaluation model by the objective function, to finally obtain a trained policy evaluation model.

The content of the target strategy condition can be input into a trained strategy evaluation model to obtain an evaluation result corresponding to the target strategy condition, if the evaluation result indicates that the prediction of the target strategy condition accords with the cognitive condition corresponding to expert experience, the target strategy condition can be added into the service wind control strategy corresponding to the historical service data, and if the evaluation result indicates that the prediction of the target strategy condition does not accord with the cognitive condition corresponding to expert experience, the target strategy condition is refused to be added into the service wind control strategy corresponding to the historical service data. By means of the method, the stock wind control strategy can be upgraded and iterated in an intelligent mode, on one hand, the accuracy of the risk strategy can be improved through a large amount of data support, and the user experience is optimized; on the other hand, unnecessary resource expenditure such as follow-up risk management and control, user complaints, manual restriction solving and the like caused by misaudit of the wind control strategy can be reduced.

The embodiment of the specification provides an updating method of a wind control strategy, which comprises the steps of performing equal frequency cutting on a range of characteristic values corresponding to first characteristics in historical service data by acquiring the historical service data containing the first characteristics, obtaining numerical values of a plurality of different cutting points corresponding to the first characteristics, then constructing two mutually exclusive strategy conditions based on the numerical values of each cutting point, acquiring risk leakage rate and/or risk management rate corresponding to each strategy condition, then acquiring target strategy conditions that the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, finally determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions into the service wind control strategy corresponding to the historical service data based on the evaluation result; on the other hand, unnecessary resource expenditure such as follow-up risk management and control, user complaints, manual restriction solving and the like caused by misaudit of the wind control strategy can be reduced. In addition, the new conditions are recommended through the existing strategy, so that energy consumption caused by disturbance of a plurality of strategies is reduced, overall risk management and control experience is optimized, the overall mode is based on big data statistical information, the digital expression of the appearance of the updated strategy new conditions is calculated, abstract artificial cognition is quantized, the advantages of artificial experience and data statistics are fused with the new strategy conditions, namely the overfitting problem caused by excessive dependence on statistical indexes is reduced, and artificial experience is intelligently introduced to carry out risk judgment.

Example two

As shown in fig. 2, the embodiment of the present disclosure provides a method for updating a wind control policy, where an execution subject of the method may be a terminal device or a server, where the terminal device may be a mobile terminal device such as a mobile phone, a tablet computer, or a computer device such as a notebook computer or a desktop computer, or may also be an IoT device (specifically, such as a smart watch, a vehicle-mounted device, or the like), and where the server may be a separate server, or may be a server cluster formed by a plurality of servers, and the server may be a background server such as a financial service or an online shopping service, or may be a background server of a certain application program, or the like. In this embodiment, the execution subject is taken as a server for example for detailed description, and for the case that the execution subject is a terminal device, the following processing of the case of the server may be referred to, and will not be described herein. The method specifically comprises the following steps:

in step S202, a plurality of different policy condition samples and an identification of each policy condition sample are acquired.

In implementation, a plurality of different policy condition samples and an identifier of each policy condition sample may be acquired in a plurality of different manners, for example, for a specified target service, the server may record relevant data (i.e., service data) generated during the process of executing the target service by the user each time the user executes the target service, where the relevant data may include service data such as sequence data of operation behaviors of the user, relevant data of the target service, relevant data input by the user, relevant data selected by the user, and relevant data provided by a server of the target service. After a certain period of data accumulation (such as a month before the current time, a month before the current time or a year before the current time), a certain amount of historical service data with a certain data amount can be recorded in the server, the historical service data with a certain amount and a certain data amount or part of the historical service data in the certain amount and the certain data amount can be analyzed, the characteristics contained in the historical service data can be obtained based on the obtained analysis result, and corresponding policy condition samples can be determined through the obtained characteristics, and specific processing modes can be referred to the related contents and are not repeated herein. In addition, the identification of each policy condition sample may also be obtained, so that a plurality of different policy condition samples and the identification of each policy condition sample may be obtained in the above manner.

Or, a plurality of different policy condition samples and the identifier of each policy condition sample may be directly obtained from a specified database, or, a certain amount of historical service data with a certain data amount may be obtained from a preset database, the historical service data may be analyzed, the characteristics included in the historical service data may be obtained based on the obtained analysis result, and the corresponding plurality of different policy condition samples may be determined through the obtained characteristics, in addition, the identifier of each policy condition sample may be obtained, or the like, which may be specifically set according to the actual situation, and the embodiment of the present specification does not limit the present specification.

In step S204, an identifier of one or more different policy condition samples is selected from the identifiers of the plurality of different policy condition samples, and masking processing is performed on the identifier of the selected policy condition sample.

In implementation, for example, as shown in fig. 3, a plurality of different policy condition samples corresponding to "historical average amount", "whether the transaction counterpart is complained" … "and the like, and identifications of the policy condition samples (i.e., names of the policy condition samples in fig. 3, i.e., historical average amount, whether the transaction counterpart is complained, … transaction time and the like) may be obtained, then, from the identifications of the policy condition samples, identifications of one policy condition sample may be selected, as shown in fig. 3, i.e., whether the transaction counterpart is complained, and the identifications of the selected policy condition samples may be masked, i.e., the identifications of the selected policy condition samples may be masked by a MASK layer in fig. 3, so as to MASK whether the identifications of the selected policy condition samples are complained.

In step S206, a plurality of different policy condition samples and identifications of unselected policy condition samples are input into a policy evaluation model, so as to obtain prediction information for the identifications of selected policy condition samples.

In implementation, a plurality of different policy condition samples and corresponding identifications of the policy condition samples may be spliced to obtain input data, the input data may be input into a policy evaluation model, and the identifications of the policy condition samples processed by the mask in the input data may be predicted by the policy evaluation model to obtain predicted identifications (i.e., prediction information).

In step S208, based on the above prediction information and the identification of the selected policy condition sample, the model parameters of the policy evaluation model are adjusted to perform model training on the policy evaluation model, thereby obtaining a trained policy evaluation model.

The policy evaluation model may be constructed by using a plurality of different algorithms or models, and in practical application, the policy evaluation model may be a model constructed based on a transducer, where the policy evaluation model may include one or more transducer modules, and if the policy evaluation model includes a plurality of transducer modules, a structure between the plurality of transducer modules may be set according to a practical situation. Alternatively, the policy evaluation model may be a model constructed based on LSMT, and may specifically be set according to actual situations, which is not limited in the embodiment of the present specification.

In implementation, the corresponding loss function or objective function may be preset according to the actual situation. Corresponding loss information can be calculated through a preset loss function according to the prediction information and the identification of the selected strategy condition sample, and the model parameters of the strategy evaluation model can be adjusted through the loss information to obtain an adjusted strategy evaluation model. Then, the model parameters of the policy evaluation model may be adjusted by the processing of the above steps S202 to S208 until the loss function converges or the objective function is satisfied, and finally the trained policy evaluation model is obtained.

In practical applications, the above-mentioned process of adjusting the model parameters of the policy evaluation model based on the above-mentioned prediction information and the identification of the selected policy condition sample may be various, and an optional processing manner is provided below, which may specifically include the following processing in step A2 and step A4.

In step A2, a sample similarity between the prediction information and the identity of the selected policy condition sample is determined.

In implementation, the above-mentioned prediction information and the identification of the selected policy condition sample may be vectorized to obtain a vector corresponding to the above-mentioned prediction information and a vector corresponding to the identification of the selected policy condition sample, and the distance between the vector corresponding to the above-mentioned prediction information and the vector corresponding to the identification of the selected policy condition sample may be calculated by means of cosine distance or euclidean distance, etc., so as to obtain a distance value that may be used as a sample similarity between the above-mentioned prediction information and the identification of the selected policy condition sample.

It should be noted that, in addition to determining the sample similarity between the prediction information and the identifier of the selected policy condition sample in the above manner, the sample similarity may also be determined in other manners, for example, the sample similarity may be determined in a manner of semantic matching (that is, determining the semantic information corresponding to the prediction information, determining the semantic information corresponding to the identifier of the selected policy condition sample, and performing a matching calculation on the two semantic information).

In step A4, model parameters of the policy evaluation model are adjusted based on the determined sample similarity.

In the implementation, in the practical application, the greater the determined sample similarity is, the more similar the prediction information is to the identification of the selected policy condition sample, the threshold may be preset according to the practical situation, and the model parameters of the policy evaluation model may be adjusted based on the threshold and the determined sample similarity.

In step S210, historical service data including the first feature is obtained, and the range of feature values corresponding to the first feature in the historical service data is subjected to equal-frequency cutting to obtain the numerical values of a plurality of different cutting points corresponding to the first feature.

In step S212, for the value of any one of the cut points, one of two mutually exclusive policy conditions is constructed based on the value greater than or equal to the cut point; and constructing the other strategy condition of the two mutually exclusive strategy conditions based on the numerical value smaller than or equal to the cutting point.

In step S214, a risk leakage rate and/or a risk management rate corresponding to each policy condition are obtained.

In step S216, a target policy condition that the risk leakage rate and/or the risk management rate satisfy a preset policy screening condition is obtained from the constructed policy conditions, where the policy screening condition is constructed based on the risk leakage rate and/or the risk management rate.

In step S218, the target policy condition is input into a pre-trained policy evaluation model, and a prediction identifier corresponding to the target policy condition is obtained.

In step S220, a similarity between the predicted identifier corresponding to the target policy condition and the identifier of the first feature is determined, and an evaluation result corresponding to the target policy condition is determined based on the determined similarity.

The evaluation result comprises a grading value for adding target strategy conditions to the business wind control strategy, wherein the target strategy conditions comprise a plurality of target strategy conditions.

In practice, the determined similarity may be used as a score in the evaluation result.

In step S222, sorting from large to small is performed on the plurality of target policy conditions based on the score value corresponding to each target policy condition, so as to obtain a plurality of sorted target policy conditions.

In step S224, a preset number of target policy conditions in front of the ordered target policy conditions are obtained, and the obtained preset number of target policy conditions are added to the service wind control policy corresponding to the historical service data.

The preset number may be set according to practical situations, such as 1, 5, or 10.

The process of determining whether to add the target policy condition to the service wind control policy corresponding to the historical service data based on the evaluation result in the above step S108 may be variously performed by the above steps S222 and S224, or may be performed by the following steps B2 and B4.

In step B2, one or more different target policy conditions with score values higher than a preset score threshold are obtained from the plurality of target policy conditions based on the score value corresponding to each target policy condition.

The preset scoring threshold may be set according to practical situations, for example, 0.9 or 0.95.

In step B4, the obtained one or more different target policy conditions are added to the service wind control policy corresponding to the historical service data.

Example III

The following provides a detailed description of an update method of a wind control policy in an embodiment of the present disclosure in connection with a specific application scenario, where an identifier of a policy condition sample may be a name of the policy condition sample, prediction information for an identifier of a selected policy condition sample may be a prediction identifier for an identifier of the selected policy condition sample, and an evaluation result may be a score value.

As shown in fig. 4, the embodiment of the present disclosure provides a method for updating a wind control policy, where an execution subject of the method may be a terminal device or a server, where the terminal device may be a mobile terminal device such as a mobile phone, a tablet computer, or a computer device such as a notebook computer or a desktop computer, or may also be an IoT device (specifically, such as a smart watch, a vehicle-mounted device, or the like), and where the server may be a separate server, or may be a server cluster formed by a plurality of servers, and the server may be a background server such as a financial service or an online shopping service, or may be a background server of a certain application program, or the like. In this embodiment, the execution subject is taken as a server for example for detailed description, and for the case that the execution subject is a terminal device, the following processing of the case of the server may be referred to, and will not be described herein. The method specifically comprises the following steps:

In step S402, a plurality of different policy condition samples and the name of each policy condition sample are acquired.

In step S404, one or more different names of policy condition samples are selected from the names of the plurality of different policy condition samples, and mask processing is performed on the selected names of the policy condition samples.

In step S406, names of a plurality of different policy condition samples and unselected policy condition samples are input into the policy evaluation model, and a prediction identifier for the name of the selected policy condition sample is obtained.

In step S408, a sample similarity between the prediction identity and the name of the selected policy condition sample is determined.

In step S410, based on the determined sample similarity, model parameters of the policy evaluation model constructed based on the transducer are adjusted to perform model training on the policy evaluation model, thereby obtaining a trained policy evaluation model.

In step S412, historical service data including the first feature is obtained, and the range of the feature value corresponding to the first feature in the historical service data is subjected to equal-frequency cutting to obtain the numerical values of a plurality of different cutting points corresponding to the first feature.

In implementation, as shown in fig. 5, for a specified target service, the server may record service data generated during the user's execution of the target service each time the user executes the target service. After a certain period of data accumulation, a certain amount of historical service data with a certain data volume can be recorded in the server. The historical service data with a certain quantity and a certain data quantity or part of the historical service data can be obtained, the historical service data with a certain quantity and a certain data quantity or part of the historical service data can be analyzed, the characteristics contained in the historical service data can be obtained through analysis results, and the range of the characteristic value corresponding to each characteristic can be determined. The features contained in the historical service data can be traversed, and for any traversed feature (namely, the first feature), the range of feature values corresponding to the first feature in the historical service data can be subjected to equal-frequency cutting, for example, 100 times of cutting can be performed, and finally, the values of a plurality of different cutting points corresponding to the first feature can be obtained, namely, the values of 100 cutting points can be obtained.

In step S414, for any value of the cut points, one of two mutually exclusive policy conditions is constructed based on the value greater than or equal to the cut point; and constructing the other strategy condition of the two mutually exclusive strategy conditions based on the numerical value smaller than or equal to the cutting point.

In implementation, as shown in fig. 5, the values of 100 cut points of the first feature may be traversed, where each cut point value may produce two mutually exclusive policy conditions, one of which is a policy condition constructed based on a value greater than or equal to the cut point, and the other of which is a policy condition constructed based on a value less than or equal to the cut point.

In step S416, a risk leakage rate and/or a risk management rate corresponding to each policy condition are obtained.

In step S418, a target policy condition that the risk leakage rate and/or the risk management rate satisfy a preset policy screening condition is obtained from the constructed policy conditions, where the policy screening condition is constructed based on the risk leakage rate and/or the risk management rate.

In implementation, as shown in fig. 5, a target policy condition that the risk leakage rate and/or the risk management rate meet a preset policy screening condition may be obtained from the constructed policy conditions, and the target policy condition may be added to the candidate list. If the two policy conditions do not meet the preset policy filtering conditions, the numerical value of the next cutting point of the first feature may be traversed through step S412.

In step S420, the target policy condition is input into a pre-trained policy evaluation model, so as to obtain a prediction identifier corresponding to the target policy condition.

In step S422, a similarity between the prediction identifier corresponding to the target policy condition and the name of the first feature is determined, and a score value corresponding to the target policy condition is determined based on the determined similarity, where the score value is used to characterize a probability that the target policy condition can be added to the service wind control policy.

Wherein, the higher the score value, the higher the probability of adding the target policy condition to the business wind control policy is, and the target policy condition can comprise a plurality of target policy conditions.

In step S424, sorting from large to small is performed on the plurality of target policy conditions based on the score value corresponding to each target policy condition, so as to obtain a plurality of sorted target policy conditions.

In step S426, a preset number of target policy conditions in front of the ordered target policy conditions are obtained, and the obtained preset number of target policy conditions are added to the service wind control policy corresponding to the historical service data.

In practical application, if the traversing of the values of the 100 cut points of the first feature is completed, the traversing of the features after the first feature in the features included in the history service data may be continued, and the processing from step S412 to step S426 may be performed until the traversing of the features included in the history service data is completed.

Example IV

The above method for updating the wind control policy provided in the embodiment of the present disclosure is based on the same concept, and the embodiment of the present disclosure further provides an apparatus for updating the wind control policy, as shown in fig. 6.

The updating device of the wind control strategy comprises: a feature cutting module 601, a policy construction module 602, a screening module 603 and a wind control policy updating module 604, wherein:

the feature cutting module 601 obtains historical service data containing a first feature, and performs equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature;

the policy construction module 602 constructs two mutually exclusive policy conditions based on the value of each cutting point, and obtains the risk leakage rate and/or the risk management rate corresponding to each policy condition;

the screening module 603 acquires target policy conditions that the risk leakage rate and/or the risk management rate meet preset policy screening conditions from the constructed policy conditions, wherein the policy screening conditions are constructed based on the risk leakage rate and/or the risk management rate;

the wind control policy updating module 604 determines, based on the target policy condition, an evaluation result corresponding to the target policy condition through a pre-trained policy evaluation model, and determines whether to add the target policy condition to the service wind control policy corresponding to the historical service data based on the evaluation result.

In the embodiment of the present disclosure, the policy building module 602 includes:

the first construction unit constructs one strategy condition of the two mutually exclusive strategy conditions based on the numerical value which is larger than or equal to the cutting point;

and the second construction unit constructs the other strategy condition of the two mutually exclusive strategy conditions based on the numerical value smaller than or equal to the cutting point.

In an embodiment of the present disclosure, the apparatus further includes:

the sample acquisition module acquires a plurality of different strategy condition samples and identifications of each strategy condition sample;

the mask processing module is used for selecting one or more identifications of different strategy condition samples from the identifications of the plurality of different strategy condition samples and performing mask processing on the identifications of the selected strategy condition samples;

the prediction module is used for inputting the identifiers of the plurality of different strategy condition samples and the unselected strategy condition samples into the strategy evaluation model to obtain the prediction information aiming at the identifiers of the selected strategy condition samples;

and the training module is used for adjusting the model parameters of the strategy evaluation model based on the prediction information and the identification of the selected strategy condition sample so as to carry out model training on the strategy evaluation model and obtain a trained strategy evaluation model.

In an embodiment of the present disclosure, the training module includes:

a similarity determining unit for determining a sample similarity between the prediction information and the identification of the selected policy condition sample;

and the parameter adjustment unit is used for adjusting the model parameters of the strategy evaluation model based on the determined sample similarity.

In the embodiment of the present specification, the policy evaluation model is a model constructed based on a transducer, or the policy evaluation model is a model constructed based on LSMT.

In this embodiment of the present disclosure, the evaluation result includes a score value that adds the target policy condition to the service wind control policy, where the target policy condition includes a plurality of target policy conditions, and the wind control policy updating module 604 includes:

the selecting unit is used for acquiring one or more different target strategy conditions with the grading values higher than a preset grading threshold value from the target strategy conditions based on the grading values corresponding to the target strategy conditions;

and the first wind control strategy updating unit is used for adding the acquired one or more different target strategy conditions into the service wind control strategy.

The sorting unit is used for sorting the plurality of target strategy conditions from large to small based on the scoring value corresponding to each target strategy condition to obtain a plurality of sorted target strategy conditions;

the second wind control strategy updating unit acquires a preset number of target strategy conditions which are sequenced in front from the sequenced target strategy conditions, and adds the acquired preset number of target strategy conditions into the service wind control strategy.

In the embodiment of the present disclosure, the wind control policy updating module 604 includes:

the prediction unit is used for inputting the target strategy conditions into a pre-trained strategy evaluation model to obtain a prediction identifier corresponding to the target strategy conditions;

and the evaluation unit is used for determining the similarity between the predicted identifier corresponding to the target strategy condition and the identifier of the first feature, and determining an evaluation result corresponding to the target strategy condition based on the determined similarity.

The embodiment of the specification provides an updating device of a wind control strategy, which is characterized in that through obtaining historical service data containing first characteristics, equal-frequency cutting is carried out on the range of characteristic values corresponding to the first characteristics in the historical service data to obtain the numerical values of a plurality of different cutting points corresponding to the first characteristics, then, two mutually exclusive strategy conditions can be constructed based on the numerical values of each cutting point, risk leakage rate and/or risk management rate corresponding to each strategy condition are obtained, then, target strategy conditions that the risk leakage rate and/or the risk management rate meet preset strategy screening conditions can be obtained from the constructed strategy conditions, finally, based on the target strategy conditions, an evaluation result corresponding to the target strategy conditions is determined through a pre-trained strategy evaluation model, whether the target strategy conditions are added into a service wind control strategy corresponding to the historical service data is determined based on the evaluation result, so that firstly, the screening is carried out based on the statistical performance of newly added characteristics, the artificial deviation is overcome, secondly, the characteristics conforming to the conditions are screened in a mode of artificial characteristic combination of learning technology, and finally, all the learning systems can be provided with the experience of an individual experience of the existing strategy can be improved, and the experience of an individual experience is improved, and the experience of an intelligent user can be improved, on the aspect is improved, and the experience of the existing strategy is improved; on the other hand, unnecessary resource expenditure such as follow-up risk management and control, user complaints, manual restriction solving and the like caused by misaudit of the wind control strategy can be reduced. In addition, the new conditions are recommended through the existing strategy, so that energy consumption caused by disturbance of a plurality of strategies is reduced, overall risk management and control experience is optimized, the overall mode is based on big data statistical information, the digital expression of the appearance of the updated strategy new conditions is calculated, abstract artificial cognition is quantized, the advantages of artificial experience and data statistics are fused with the new strategy conditions, namely the overfitting problem caused by excessive dependence on statistical indexes is reduced, and artificial experience is intelligently introduced to carry out risk judgment.

Example five

The above device for updating the wind control policy provided in the embodiment of the present disclosure further provides an apparatus for updating the wind control policy based on the same concept, as shown in fig. 7.

The update device of the wind control policy may provide a terminal device or a server for the above embodiment.

The update device of the wind control strategy may have a relatively large difference due to different configurations or performances, and may include one or more processors 701 and a memory 702, where one or more stored applications or data may be stored in the memory 702. Wherein the memory 702 may be transient storage or persistent storage. The application program stored in memory 702 may include one or more modules (not shown in the figures), each of which may include a series of computer-executable instructions in the update device to the wind control strategy. Still further, the processor 701 may be configured to communicate with the memory 702 and execute a series of computer executable instructions in the memory 702 on an update device of the wind control strategy. The updating device of the wind control strategy may also include one or more power supplies 703, one or more wired or wireless network interfaces 704, one or more input/output interfaces 705, and one or more keyboards 706.

In particular, in this embodiment, the update device of the wind control strategy includes a memory, and one or more programs, where the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the update device of the wind control strategy, and executing the one or more programs by the one or more processors includes computer executable instructions for:

acquiring historical service data containing a first feature, and performing equal-frequency cutting on a range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature;

based on the numerical value of each cutting point, constructing two mutually exclusive strategy conditions, and acquiring a risk leakage rate and/or a risk management rate corresponding to each strategy condition;

acquiring target strategy conditions of which the risk leakage rate and/or risk management rate meet preset strategy screening conditions from constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate;

Based on the target strategy conditions, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model, and determining whether to add the target strategy conditions to the service wind control strategies corresponding to the historical service data or not based on the evaluation result.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for the update device embodiment of the wind control strategy, the description is relatively simple, as it is substantially similar to the method embodiment, as relevant see the section of the method embodiment.

The embodiment of the specification provides updating equipment of a wind control strategy, which is characterized in that by acquiring historical service data containing first characteristics, equal-frequency cutting is carried out on the range of characteristic values corresponding to the first characteristics in the historical service data to obtain the numerical values of a plurality of different cutting points corresponding to the first characteristics, then, two mutually exclusive strategy conditions can be constructed based on the numerical values of each cutting point, risk leakage rate and/or risk management rate corresponding to each strategy condition are acquired, then, target strategy conditions that the risk leakage rate and/or the risk management rate meet preset strategy screening conditions can be acquired from the constructed strategy conditions, finally, an evaluation result corresponding to the target strategy conditions can be determined through a pre-trained strategy evaluation model, and whether the target strategy conditions are added into the service wind control strategy corresponding to the historical service data is determined based on the evaluation result; on the other hand, unnecessary resource expenditure such as follow-up risk management and control, user complaints, manual restriction solving and the like caused by misaudit of the wind control strategy can be reduced. In addition, the new conditions are recommended through the existing strategy, so that energy consumption caused by disturbance of a plurality of strategies is reduced, overall risk management and control experience is optimized, the overall mode is based on big data statistical information, the digital expression of the appearance of the updated strategy new conditions is calculated, abstract artificial cognition is quantized, the advantages of artificial experience and data statistics are fused with the new strategy conditions, namely the overfitting problem caused by excessive dependence on statistical indexes is reduced, and artificial experience is intelligently introduced to carry out risk judgment.

Example six

Further, based on the method shown in fig. 1 to 5, one or more embodiments of the present disclosure further provide a storage medium, which is used to store computer executable instruction information, and in a specific embodiment, the storage medium may be a U disc, an optical disc, a hard disk, etc., where the computer executable instruction information stored in the storage medium can implement the following flow when executed by a processor:

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for one of the above-described storage medium embodiments, since it is substantially similar to the method embodiment, the description is relatively simple, and reference is made to the description of the method embodiment for relevant points.

The embodiment of the specification provides a storage medium, through obtaining historical service data containing first characteristics, carrying out equal frequency cutting on a range of characteristic values corresponding to the first characteristics in the historical service data, obtaining numerical values of a plurality of different cutting points corresponding to the first characteristics, then, constructing two mutually exclusive strategy conditions based on the numerical values of each cutting point, obtaining risk leakage rate and/or risk management rate corresponding to each strategy condition, afterwards, obtaining target strategy conditions that the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, finally, determining an evaluation result corresponding to the target strategy conditions through a pre-trained strategy evaluation model based on the target strategy conditions, determining whether to add the target strategy conditions into a service wind control strategy corresponding to the historical service data or not based on the evaluation result, so that firstly, screening is carried out based on the statistical performance of the newly-added characteristics, the deviation of manual experience is overcome, secondly, all the characteristics conforming to the conditions are screened in a mode of learning manual characteristic combination through NLP technology, and finally, the actual strategy quality of the user can be provided with high-level cognitive learning experience through the existing strategy can be improved, on the one hand, and the accuracy of the user experience can be improved through the iterative strategy is improved; on the other hand, unnecessary resource expenditure such as follow-up risk management and control, user complaints, manual restriction solving and the like caused by misaudit of the wind control strategy can be reduced. In addition, the new conditions are recommended through the existing strategy, so that energy consumption caused by disturbance of a plurality of strategies is reduced, overall risk management and control experience is optimized, the overall mode is based on big data statistical information, the digital expression of the appearance of the updated strategy new conditions is calculated, abstract artificial cognition is quantized, the advantages of artificial experience and data statistics are fused with the new strategy conditions, namely the overfitting problem caused by excessive dependence on statistical indexes is reduced, and artificial experience is intelligently introduced to carry out risk judgment.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

In the 90 s of the 20 th century, improvements to one technology could clearly be distinguished as improvements in hardware (e.g., improvements to circuit structures such as diodes, transistors, switches, etc.) or software (improvements to the process flow). However, with the development of technology, many improvements of the current method flows can be regarded as direct improvements of hardware circuit structures. Designers almost always obtain corresponding hardware circuit structures by programming improved method flows into hardware circuits. Therefore, an improvement of a method flow cannot be said to be realized by a hardware entity module. For example, a programmable logic device (Programmable Logic Device, PLD) (e.g., field programmable gate array (Field Programmable Gate Array, FPGA)) is an integrated circuit whose logic function is determined by the programming of the device by a user. A designer programs to "integrate" a digital system onto a PLD without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Moreover, nowadays, instead of manually manufacturing integrated circuit chips, such programming is mostly implemented by using "logic compiler" software, which is similar to the software compiler used in program development and writing, and the original code before the compiling is also written in a specific programming language, which is called hardware description language (Hardware Description Language, HDL), but not just one of the hdds, but a plurality of kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), lava, lola, myHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog are currently most commonly used. It will also be apparent to those skilled in the art that a hardware circuit implementing the logic method flow can be readily obtained by merely slightly programming the method flow into an integrated circuit using several of the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable logic controllers, and embedded microcontrollers, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, atmel AT91SAM, microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic of the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller in a pure computer readable program code, it is well possible to implement the same functionality by logically programming the method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Such a controller may thus be regarded as a kind of hardware component, and means for performing various functions included therein may also be regarded as structures within the hardware component. Or even means for achieving the various functions may be regarded as either software modules implementing the methods or structures within hardware components.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing one or more embodiments of the present description.

It will be appreciated by those skilled in the art that embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, one or more embodiments of the present description can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

Embodiments of the present description are described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable fraud case serial-to-parallel device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable fraud case serial-to-parallel device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

One or more embodiments of the present specification may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the present description may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

The foregoing is merely exemplary of the present disclosure and is not intended to limit the present disclosure. Various modifications and alterations to this specification will become apparent to those skilled in the art. Any modifications, equivalent substitutions, improvements, or the like, which are within the spirit and principles of the present description, are intended to be included within the scope of the claims of the present description.

Claims

1. A method of updating a wind control strategy, the method comprising:

2. The method of claim 1, wherein the constructing two mutually exclusive policy conditions based on the value of each cut point comprises:

Constructing one of the two mutually exclusive strategy conditions based on the numerical value which is greater than or equal to the cutting point;

and constructing the other strategy condition of the two mutually exclusive strategy conditions based on the numerical value smaller than or equal to the cutting point.

3. The method of claim 2, the method further comprising:

acquiring a plurality of different strategy condition samples and an identification of each strategy condition sample;

selecting one or more different identifications of the policy condition samples from the identifications of the plurality of different policy condition samples, and masking the selected identifications of the policy condition samples;

inputting the identifications of the different strategy condition samples and the unselected strategy condition samples into the strategy evaluation model to obtain the prediction information aiming at the identifications of the selected strategy condition samples;

and adjusting model parameters of the strategy evaluation model based on the prediction information and the identification of the selected strategy condition sample so as to perform model training on the strategy evaluation model and obtain a trained strategy evaluation model.

4. A method according to claim 3, said adjusting model parameters of the policy evaluation model based on the prediction information and the identification of the selected policy condition samples, comprising:

Determining sample similarity between the prediction information and the identification of the selected policy condition samples;

and adjusting model parameters of the strategy evaluation model based on the determined sample similarity.

5. The method of claim 3, wherein the policy evaluation model is a model constructed based on a transducer, or wherein the policy evaluation model is a model constructed based on LSMT.

6. The method according to any one of claims 1-5, wherein the evaluation result comprises a scoring value for adding the target policy condition to the business wind control policy, wherein the target policy condition comprises a plurality of,

the determining whether to add the target policy condition to the service wind control policy corresponding to the historical service data based on the evaluation result includes:

based on the scoring value corresponding to each target strategy condition, acquiring one or more different target strategy conditions with scoring values higher than a preset scoring threshold value from a plurality of target strategy conditions;

and adding the acquired one or more different target strategy conditions to the business wind control strategy.

7. The method according to any one of claims 1-5, wherein the evaluation result comprises a scoring value for adding the target policy condition to the business wind control policy, wherein the target policy condition comprises a plurality of,

based on the scoring value corresponding to each target strategy condition, sorting the plurality of target strategy conditions from large to small to obtain a plurality of sorted target strategy conditions;

acquiring a preset number of target strategy conditions which are sequenced in front from a plurality of target strategy conditions which are sequenced in front, and adding the acquired preset number of target strategy conditions into the business wind control strategy.

8. The method according to claim 1, wherein the determining, based on the target policy condition, the evaluation result corresponding to the target policy condition through a pre-trained policy evaluation model includes:

inputting the target strategy condition into a pre-trained strategy evaluation model to obtain a prediction identifier corresponding to the target strategy condition;

and determining the similarity between the predicted identifier corresponding to the target strategy condition and the identifier of the first feature, and determining an evaluation result corresponding to the target strategy condition based on the determined similarity.

9. An apparatus for updating a wind control strategy, the apparatus comprising:

The feature cutting module is used for obtaining historical service data containing a first feature, and performing equal-frequency cutting on the range of feature values corresponding to the first feature in the historical service data to obtain numerical values of a plurality of different cutting points corresponding to the first feature;

the strategy construction module is used for constructing two mutually exclusive strategy conditions based on the numerical value of each cutting point and acquiring the risk leakage rate and/or the risk management rate corresponding to each strategy condition;

the screening module is used for acquiring target strategy conditions that the risk leakage rate and/or the risk management rate meet preset strategy screening conditions from the constructed strategy conditions, wherein the strategy screening conditions are constructed based on the risk leakage rate and/or the risk management rate;

and the wind control strategy updating module is used for determining an evaluation result corresponding to the target strategy condition through a pre-trained strategy evaluation model based on the target strategy condition, and determining whether to add the target strategy condition into the service wind control strategy corresponding to the historical service data based on the evaluation result.

10. An update apparatus of a wind control policy, the update apparatus of a wind control policy comprising:

a processor; and

a memory arranged to store computer executable instructions that, when executed, cause the processor to: