CN116739759A

CN116739759A - Asset and fund matching method, device and equipment based on order risk identification

Info

Publication number: CN116739759A
Application number: CN202310866252.1A
Authority: CN
Inventors: 赵薇
Original assignee: Shenzhen Lexin Software Technology Co Ltd
Current assignee: Shenzhen Lexin Software Technology Co Ltd
Priority date: 2023-07-14
Filing date: 2023-07-14
Publication date: 2023-09-12

Abstract

The invention discloses an asset fund matching method, device and equipment based on order risk identification, wherein the method comprises the steps of sequentially carrying out risk identification based on a two-class prediction model set and a decision tree strategy model set after acquiring a user asset order data set to obtain a corresponding order risk identification result set; and acquiring target user asset order data corresponding to the order risk identification result with the first preset type identification result in the order risk identification result set, and acquiring corresponding fund party matching results based on a fund matching strategy. According to the method and the device, the target user asset order data with the first preset type identification result can be screened to participate in final capital asset matching after the two-time risk assessment of the user asset order data set by adopting the two-time classification prediction model set and the decision tree strategy model set, so that the data utilization rate of the asset order is improved, and the problem that the user asset order data is intercepted by mistake is avoided due to the adoption of secondary risk identification.

Description

Asset and fund matching method, device and equipment based on order risk identification

Technical Field

The invention relates to the technical field of intelligent decision making of artificial intelligence, in particular to an asset fund matching method, device and equipment based on order risk identification.

Background

In the field of internet finance, fund asset matching is a common data processing flow of a fund asset matching platform. Wherein the property is formed by a loan application order (e.g., including user information, order amount, account period, and interest rate) issued by a user to an internet finance company; the fund party is a financial institution or bank with which the fund asset matching platform cooperates and which can provide funds. The fund property matching is to match the loan application order corresponding to the property to the most suitable fund party, and the subsequent links can be carried out after the loan application order is checked by the fund party.

At present, when the fund property matching platform performs fund property matching, the risk level of the loan application order is determined after the risk assessment is performed on the loan application order once by the air control strategy, and then if the risk level of the loan application order is determined to be high risk level or medium risk level, the loan application order is directly intercepted and the subsequent fund party matching is not participated.

However, based on the current manner of determining the risk level of a loan application order by adopting a risk assessment of a wind control strategy line once and determining whether to intercept the order, the following defects exist:

1) The situation of misjudging the risk level of the order exists, so that the loan application order is intercepted by mistake;

2) The loan application orders corresponding to the risk levels are potential matching promotion orders, and the fund property matching platform is used for uniformly intercepting the potential matching promotion orders and does not evaluate other indexes of the loan application orders further, so that the success rate of successful fund property matching can be finally promoted by the fund property matching platform, and the utilization rate of the loan application orders serving as the assets in the fund property matching platform is low.

Disclosure of Invention

The embodiment of the invention provides a method, a device and equipment for matching assets based on order risk identification, which aim to solve the problem that in the prior art, a loan application order is intercepted by mistake due to the fact that a risk level of the asset matching platform adopts a wind control strategy to conduct one-time risk assessment to determine whether the order is intercepted or not.

In a first aspect, an embodiment of the present invention provides an asset funds matching method based on order risk identification, including:

responding to an asset fund matching instruction, and acquiring a user asset order data set corresponding to the asset fund matching instruction;

Acquiring a prediction result set corresponding to the user asset order data set based on a pre-trained two-classification prediction model set; wherein each user asset order data in the user asset order dataset corresponds to a prediction result;

acquiring the corresponding supplementary user characteristics of each predicted result in the predicted result set to update each predicted result, and obtaining an updated predicted result set;

acquiring an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree strategy model set; wherein each updated prediction result in the updated prediction result set corresponds to an order risk identification result;

acquiring an order risk recognition result with a first preset type recognition result in the order risk recognition result set to form a target order risk recognition result set;

and acquiring target user asset order data corresponding to each target order risk identification result in the target order risk identification result set respectively, and acquiring fund party matching results corresponding to each target user asset order data based on a preset fund matching strategy.

In a second aspect, an embodiment of the present invention further provides an asset funds matching device based on order risk identification, including:

An order data acquisition unit, which is used for responding to the asset fund matching instruction and acquiring a user asset order data set corresponding to the asset fund matching instruction;

the two-classification prediction unit is used for acquiring a prediction result set corresponding to the user asset order data set based on a pre-trained two-classification prediction model set; wherein each user asset order data in the user asset order dataset corresponds to a prediction result;

the prediction result updating unit is used for acquiring the complementary user characteristics corresponding to each prediction result in the prediction result set to update each prediction result, so as to obtain an updated prediction result set;

the decision unit is used for acquiring an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree strategy model set; wherein each updated prediction result in the updated prediction result set corresponds to an order risk identification result;

the recognition result screening unit is used for acquiring the order risk recognition results with the first preset type recognition results in the order risk recognition result set to form a target order risk recognition result set;

and the asset and fund matching unit is used for acquiring the asset order data of the target users, which correspond to the target order risk identification results in the target order risk identification result set, and acquiring fund party matching results corresponding to the asset order data of the target users based on a preset fund matching strategy.

In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory and a processor, where the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

In a fourth aspect, embodiments of the present invention also provide a computer readable storage medium storing a computer program comprising program instructions which, when executed by a processor, implement the method of the first aspect.

The embodiment of the invention provides an asset fund matching method, device and equipment based on order risk identification, wherein the method comprises the following steps: responding to the asset fund matching instruction, and acquiring a user asset order data set corresponding to the asset fund matching instruction; acquiring a prediction result set corresponding to the user asset order data set based on a pre-trained two-classification prediction model set; acquiring the corresponding supplementary user characteristics of each predicted result in the predicted result set to update each predicted result, and obtaining an updated predicted result set; acquiring an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree strategy model set; acquiring order risk recognition results with first preset type recognition results in an order risk recognition result set to form a target order risk recognition result set; and acquiring target user asset order data corresponding to each target order risk identification result in the target order risk identification result set, and acquiring fund party matching results corresponding to each target user asset order data based on a preset fund matching strategy. According to the method and the device, the target user asset order data with the first preset type identification result can be screened to participate in final capital asset matching after the two-time risk assessment of the user asset order data set by adopting the two-time classification prediction model set and the decision tree strategy model set, so that the data utilization rate of the asset order is improved, and the problem that the user asset order data is intercepted by mistake is avoided due to the adoption of secondary risk identification.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of an application scenario of an asset fund matching method based on order risk identification according to an embodiment of the present invention;

FIG. 2 is a flow chart of an asset and funds matching method based on order risk identification provided by an embodiment of the invention;

FIG. 3 is a schematic sub-flowchart of an asset and funds matching method based on order risk identification according to an embodiment of the present invention;

FIG. 4 is a schematic block diagram of an asset funds matching device based on order risk identification provided by an embodiment of the invention;

fig. 5 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

The embodiment of the invention provides an asset fund matching method, device and equipment based on order risk identification. The asset fund matching method based on order risk identification in the embodiment of the invention is applied to a server, wherein one or more processors, a memory and one or more application programs are arranged in the server, and the one or more application programs are stored in the memory and are configured to be executed by the processor to realize the asset fund matching method based on order risk identification.

As shown in fig. 1, fig. 1 is a schematic diagram of a scenario of an asset fund matching method based on order risk identification according to an embodiment of the present invention, where the scenario of asset fund matching based on order risk identification in the embodiment of the present invention includes a server 10 and a user terminal 20, where an asset fund matching device based on order risk identification is integrated in the server 10, and a storage medium corresponding to the asset fund matching method based on order risk identification is operated to execute steps of the asset fund matching method based on order risk identification.

It may be understood that, in the server 10 in the specific application scenario of the asset and funds matching method based on order risk identification shown in fig. 1, the devices included in the server 10 do not limit the embodiments of the present invention, that is, the number of devices and the types of devices included in the specific application scenario of the asset and funds matching method based on order risk identification, or the number of devices and the types of devices included in each device do not affect the overall implementation of the technical solution in the embodiments of the present invention, and may be calculated as equivalent replacement or derivation of the claimed technical solution in the embodiments of the present invention.

The server 10 in the embodiment of the present invention is mainly used for: responding to an asset fund matching instruction, and acquiring a user asset order data set corresponding to the asset fund matching instruction; acquiring a prediction result set corresponding to the user asset order data set based on a pre-trained two-classification prediction model set; wherein each user asset order data in the user asset order dataset corresponds to a prediction result; acquiring the corresponding supplementary user characteristics of each predicted result in the predicted result set to update each predicted result, and obtaining an updated predicted result set; acquiring an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree strategy model set; wherein each updated prediction result in the updated prediction result set corresponds to an order risk identification result; acquiring an order risk recognition result with a first preset type recognition result in the order risk recognition result set to form a target order risk recognition result set; and acquiring target user asset order data corresponding to each target order risk identification result in the target order risk identification result set respectively, and acquiring fund party matching results corresponding to each target user asset order data based on a preset fund matching strategy.

It will be appreciated by those skilled in the art that the application environment shown in fig. 1 is merely an application scenario of the present invention, and is not limited to the application scenario of the present invention, and other application environments may further include more or fewer servers 10 than those shown in fig. 1, or network connection relationships of the servers 10, for example, only 1 server 10 is shown in fig. 1, and it will be appreciated that the specific application scenario of the asset fund matching method based on order risk identification may further include one or more other servers 10, which is not limited herein; memory may also be included in the server 10.

FIG. 2 is a flow chart of an asset funds matching method based on order risk identification provided by an embodiment of the invention. As shown in fig. 2, the method includes the following steps S110 to S160.

S110, responding to an asset fund matching instruction, and acquiring a user asset order data set corresponding to the asset fund matching instruction.

In this embodiment, the technical scheme is described by using a server as an execution body. For example, when the user needs to transact loan service, the user terminal fills in the specific information such as user basic information (such as user name, gender, age, unique identity number, etc.), application amount, etc. on the user interactive interface of the corresponding application program, and then generates a loan order, and the generated order can be regarded as a property order. After the property order is generated, whether the corresponding fund provider can be matched for loan fund release or not is judged after risk identification and other processing is needed (wherein the loan fund provider can be regarded as a fund party).

After the server receives the user asset order data set uploaded by the at least one user terminal and including the at least one user asset order data, all user asset order data in the user asset order data set is required to perform order risk identification and match the funding party after verification of the order risk identification.

In one embodiment, step S110 further includes:

acquiring user characteristics, sponsor characteristics and order characteristics corresponding to all user asset order data in the user asset order data set;

if the user characteristics, the sponsor characteristics and the order characteristics corresponding to the user asset order data are all determined to be non-null characteristics, the user characteristics, the sponsor characteristics and the order characteristics corresponding to the user asset order data form corresponding user asset order input data;

if the corresponding sponsor feature of the user asset order data is determined to be the null feature, the third party user data of the corresponding user asset order data is acquired to extract the third sponsor feature and update the sponsor feature, and the corresponding user asset order input data is composed of the user feature, the sponsor feature and the order feature of the corresponding user asset order data.

In this embodiment, if a piece of user asset order data 1 uploaded by a user terminal a is taken as an example, user feature 1 (such as a user name, an age, a sex, a occupation, etc. feature) and order feature 1 (such as an order account type, more specifically, an order account type with 1-3 accounts, an order account type with 4-6 accounts, an order account type with 7-12 accounts, and an order account type with more than 12 accounts) can be directly extracted from the user asset order data 1, and the historical user asset order data 1 with user feature 1 can be obtained based on a historical user asset order data set stored locally in the server, and fund party data corresponding to the historical user asset order data 1 is taken as corresponding fund party feature 1, and finally, the corresponding user asset order input data 1 is composed of the user feature 1, the fund party feature 1, and the order feature 1 corresponding to the user asset order.

Of course, when the historical user asset order data 1 with the user feature 1 is not obtained in the historical user asset order data set stored locally in the server, the third party user data 1 with the user feature 1 can be obtained in the third party server with the communication connection relation with the server based on the user feature 1, the third party user data 1 is extracted to be used as the corresponding sponsor feature 1, and finally the corresponding user asset order input data 1 is formed by the user feature 1, the sponsor feature 1 and the order feature 1 corresponding to the user asset order data 1. Therefore, the specific information of each dimension characteristic constituting the user asset order input data can be rapidly obtained through the mode.

S120, acquiring a prediction result set corresponding to the user asset order data set based on a pre-trained two-classification prediction model set; wherein each user asset order data in the user asset order dataset corresponds to a predicted outcome.

In this embodiment, a set of two-class prediction models formed by a plurality of two-class prediction models that have been trained in advance is stored locally in the server, and when each user asset order data in the user asset order data set is used as an input of one of the two-class prediction models (for example, the two-class prediction models are all two-class models for predicting the high-bad account rate of the user asset order data) to perform a prediction operation, a prediction result corresponding to each user asset order data can be obtained. Because the first risk assessment based on the user asset order data adopts a two-class model, the method has higher prediction result acquisition efficiency, and simultaneously avoids directly taking the prediction result as a final order risk assessment standard, thereby improving the order utilization rate.

In an embodiment, the set of classified prediction models includes at least one classified prediction model, the user asset order data includes at least one order ledger type, different order ledger types have different ledger durations, and the at least one classified prediction model corresponds to the ledger durations of the at least one order ledger type;

The obtaining of the prediction result set corresponding to the user asset order data set based on the pre-trained two-classification prediction model set is performed for each user asset order data of the user asset order data set:

acquiring user asset order input data of the user asset order data, and acquiring order characteristics and corresponding order accounting period types in the user asset order input data;

and determining a target two-class prediction model corresponding to the order account period type in the two-class prediction model set, and calculating the user asset order input data based on the target two-class prediction model to obtain a prediction result.

In this embodiment, taking the case that the set of two-class prediction models includes four two-class prediction models as an example, that is, the set of two-class prediction models includes a first two-class prediction model, a second two-class prediction model, a third two-class prediction model, and a fourth two-class prediction model; the first two-class prediction model is used for carrying out order risk identification on user asset order data with a first order accounting period type; the second classification prediction model is used for carrying out order risk identification on user asset order data with a second order accounting period type; the third classification prediction model is used for carrying out order risk identification on user asset order data with a third order accounting period type; the fourth classification prediction model is used for carrying out order risk identification on user asset order data with a fourth order accounting period type; the account period duration corresponding to the first order account period type is smaller than the account period duration corresponding to the second order account period type, the account period duration corresponding to the second order account period type is smaller than the account period duration corresponding to the third order account period type, and the account period duration corresponding to the third order account period type is smaller than the account period duration corresponding to the fourth order account period type;

Wherein, in step S120, when each user asset order data of the user asset order data set is used to obtain a corresponding prediction result based on the two classification prediction model sets, the following steps are performed:

if the order account period type of the user asset order input data is determined to belong to a first order account period type, calculating the user asset order input data based on the first two-class prediction model to obtain a prediction result;

if the order account period type of the user asset order input data is determined to belong to a second order account period type, calculating the user asset order input data based on the second classification prediction model to obtain a prediction result;

if the order account period type of the user asset order input data is determined to belong to a third order account period type, calculating the user asset order input data based on the third classification prediction model to obtain a prediction result;

and if the order account period type of the user asset order input data is determined to belong to a fourth order account period type, calculating the user asset order input data based on the fourth classification prediction model to obtain a prediction result.

In this embodiment, the set of classification prediction models includes at least a first classification prediction model, a second classification prediction model, a third classification prediction model, and a fourth classification prediction model for performing order risk identification on the user asset order data having the first order accounting period type, the second order accounting period type, the third order accounting period type, and the fourth order accounting period type, respectively. More specifically, an order accounting period type having 1-3 periods is taken as a first order accounting period type, an order accounting period type having 4-6 periods is taken as a second order accounting period type, an order accounting period type having 7-12 periods is taken as a third order accounting period type, and an order accounting period type having 12 or more periods is taken as a fourth order accounting period type.

When order risk identification is required for a user asset order data, determining whether an order accounting period type of the user asset order input data (for example, an order accounting period type is extracted from order features in the user asset order input data) belongs to a first order accounting period type, or belongs to a second order accounting period type, or belongs to a first order accounting period type, or belongs to a fourth order accounting period type; after determining the target order accounting period type of the user asset order input data, taking a two-class prediction model corresponding to the target order accounting period type as a target two-class prediction model (namely one of a first two-class prediction model, a second two-class prediction model, a third two-class prediction model and a fourth two-class prediction model); and finally, inputting the user asset order input data into a target classification prediction model for operation to obtain a prediction result. Therefore, different classification prediction models are adopted for classifying and predicting the user asset order data of different order account period types, so that the prediction result is more accurate.

S130, obtaining the corresponding supplementary user characteristics of each prediction result in the prediction result set to update each prediction result, and obtaining an updated prediction result set.

In this embodiment, after obtaining, in the server, a prediction result corresponding to each user asset order data in the user asset order data sets to form a prediction result set, instead of directly using the prediction result set as input data of the risk assessment model of the next time, it is also necessary to add additional user features corresponding to each prediction result to update each prediction result, so as to obtain an updated prediction result set. Because the additional user features are added to each prediction result in a targeted way before the next input to another risk assessment model, the data for carrying out risk assessment for the second time has more dimension user information, and the accuracy of the output result of the final secondary risk assessment can be effectively improved.

In one embodiment, as shown in fig. 3, step S130 includes:

s131, obtaining user asset order data corresponding to each prediction result in the prediction result set;

s132, acquiring a wind control feature set and order supplementary features corresponding to the asset order data of each user to form supplementary user features corresponding to each prediction result; the wind control feature sets corresponding to the user asset order data are obtained by inputting the user asset order data into a plurality of wind control models for calculation;

And S133, updating each prediction result by using the corresponding supplementary user characteristic of each prediction result to obtain the updated prediction result set.

In this embodiment, after obtaining, in the server, a prediction result corresponding to each piece of user asset order data in the user asset order data set, it is known that each prediction result corresponds to one piece of user asset order data. Because the server also stores a plurality of other online or offline wind control models, the wind control feature set corresponding to the user asset order data can be obtained by inputting each user asset order data into the other plurality of online or offline wind control models for operation. Also stored in the server is a historical user asset order dataset that can be analyzed for characteristics such as advance payoff rates. After the order amount of each user asset order data in the user asset order data set is obtained in the server, features such as an advance repayment rate and the order amount corresponding to each user asset order data are used as order supplement features corresponding to each user asset order data. When the wind control feature set and the order supplement feature are correspondingly added in each predicted result, namely, updating of each predicted result is completed in a data dimension increasing mode, so that each updated predicted result in the updated predicted result set has more data information.

S140, acquiring an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree strategy model set; wherein each updated prediction result in the updated prediction result set corresponds to an order risk identification result.

In this embodiment, a decision tree policy model set formed by a plurality of decision tree policy models that have been trained in advance is stored locally in the server, and when a risk identification operation is performed by using each updated prediction result in the updated prediction result set as an input of one of the decision tree policy models of the multi-decision tree policy model, an order risk identification result corresponding to each updated prediction result can be obtained. Because the decision tree model is adopted for the second risk assessment based on the updated prediction result, the order risk identification result corresponding to the original asset order data of each user can be obtained more accurately based on the second risk assessment.

In an embodiment, the set of decision tree policy models includes at least one decision tree policy model, and the at least one decision tree policy model corresponds to the at least one bi-classification prediction model;

The obtaining of the order risk recognition result set corresponding to the updated prediction result set based on the pre-trained decision tree strategy model set is performed for each updated prediction result of the updated prediction result set:

acquiring a prediction result corresponding to the updated prediction result, and taking an order account period type in user asset order input data corresponding to the prediction result as an order account period type of the updated prediction result;

and determining a target decision tree strategy model corresponding to the order account period type of the updated prediction result in the decision tree strategy model set, and calculating the updated prediction result based on the target decision tree strategy model to obtain an order risk identification result.

In this embodiment, taking the example that the set of decision tree policy models includes four decision tree policy models, including a first decision tree policy model corresponding to the first two-class prediction model, a second decision tree policy model corresponding to the second two-class prediction model, a third decision tree policy model corresponding to the third class prediction model, and a fourth decision tree policy model corresponding to the fourth class prediction model;

In step S140, when each updated prediction result of the updated prediction result set is based on the decision tree policy model set to obtain the order risk identification result, the following steps are executed:

if the order account period type of the updated prediction result is determined to belong to the first order account period type, calculating the updated prediction result based on the first decision tree strategy model to obtain an order risk identification result;

if the order account period type of the updated prediction result is determined to belong to the second order account period type, calculating the updated prediction result based on the second decision tree strategy model to obtain an order risk identification result;

if the order account period type of the updated prediction result is determined to belong to the third order account period type, calculating the updated prediction result based on the third decision tree strategy model to obtain an order risk identification result;

And if the order account period type of the updated prediction result is determined to belong to the fourth order account period type, calculating the updated prediction result based on the fourth decision tree strategy model to obtain an order risk identification result.

In this embodiment, the decision tree policy model set includes at least a first decision tree policy model, a second decision tree policy model, a third decision tree policy model, and a fourth decision tree policy model, so as to perform order risk secondary identification on the updated prediction results having the first order accounting period type, the second order accounting period type, the third order accounting period type, and the fourth order accounting period type, respectively. Likewise, after updating, the prediction result is used as a first order accounting period type if the order accounting period type with 1-3 accounting periods is used as a second order accounting period type if the order accounting period type with 4-6 accounting periods is used as a third order accounting period type if the order accounting period type with 7-12 accounting periods is used as a fourth order accounting period type if the order accounting period type with more than 12 accounting periods is used.

When secondary order risk identification is required for an updated prediction result, determining whether an order accounting period type of the updated prediction result (for example, an order accounting period type is extracted from order features in the updated prediction result) belongs to a first order accounting period type, or belongs to a second order accounting period type, or belongs to a first order accounting period type, or belongs to a fourth order accounting period type; after determining the target order accounting period type of the updated prediction result, taking a decision tree strategy model corresponding to the target order accounting period type as a target decision tree strategy model (namely one of a first decision tree strategy model, a second decision tree strategy model, a third decision tree strategy model and a fourth decision tree strategy model); and finally, inputting the updated prediction result into a target decision tree strategy model for operation to obtain an order risk identification result. Therefore, the updated prediction results of different order account period types are classified and predicted by adopting different decision tree strategy models, so that the order risk identification result is more accurate.

In one embodiment, step S140 further includes:

acquiring an initial decision tree strategy model set, and acquiring model evaluation index values obtained by each initial decision tree strategy model in the initial decision tree strategy model set based on a model test set;

if the model evaluation index value of the initial decision tree strategy model exceeds the preset evaluation index threshold value, acquiring a corresponding initial decision tree strategy model to form the decision tree strategy model set.

In this embodiment, when a decision tree policy model set is pre-built in a server, a plurality of initial decision tree policy models obtained by combining multi-dimensional features included in a prediction result after reference update are built, and then model training and testing are performed on the plurality of initial decision tree policy models by using a model training set and a model testing set in the server, so as to finally obtain model evaluation index values obtained by each initial decision tree policy model based on the model testing set. And then, selecting an initial decision tree strategy model corresponding to the model evaluation index value exceeding a preset evaluation index threshold from a plurality of initial decision tree strategy models to form the decision tree strategy model set. The decision tree model with good model performance is screened in advance, so that the obtained order risk identification result is more accurate in the subsequent practical application.

For example, taking an ROI index value (ROI, i.e., input-output ratio) of an initial decision tree strategy model as a model evaluation index value, i.e., an input-output ratio index value of the initial decision tree strategy model is to be obtained. The average proportion of profit rise generated by all orders identified as low risk results (i.e., the first preset type identification results) in the initial decision tree policy model is divided by the possible proportion of reduction in the total amount of orders submitted for all orders identified as low risk results by a model test set. If the ROI index value of the decision tree strategy model is larger, the effect of the decision tree strategy model is better.

S150, acquiring an order risk recognition result with a first preset type recognition result in the order risk recognition result set to form a target order risk recognition result set.

In this embodiment, after the order risk identification result set corresponding to the user asset order data set is obtained, the order risk identification result with the first preset type identification result therein is screened to form a target order risk identification result set. Wherein the first preset type identification result is used to identify user asset order data that does not have a significant negative profit margin. Of course, since the second preset type recognition result is further provided in the server for identifying the user asset order data with serious negative profit margin, the order risk recognition result having the second preset type recognition result in the order risk recognition result set can also be obtained in the server to form another target order risk recognition result set.

Of course, because the obtained user asset order data corresponding to each target order risk identification result in the other target order risk identification result set has serious negative profit margin, that is, the server obtains high-risk order data after two risk identifications based on the two classification prediction model sets and the decision tree strategy model sets, at this time, the server actively intercepts the user asset order data corresponding to each target order risk identification result in the other target order risk identification result set, and does not participate in later asset fund matching.

S160, acquiring target user asset order data corresponding to each target order risk identification result in the target order risk identification result set, and acquiring fund party matching results corresponding to each target user asset order data based on a preset fund matching strategy.

In this embodiment, after the target order risk identification result set is obtained in the server, a prestored fund matching policy needs to be obtained for matching the target user asset order data with the fund party, so as to achieve final asset and fund matching. By the method, the utilization rate and the matching rate of the user asset order data are effectively improved.

In an embodiment, the fund matching policy is used for matching the user asset order data according to a preset fund matching order to obtain a fund matching result, and in step S160, when determining the fund matching result for each target user asset order data of the target order risk identification result set based on the fund matching policy, the following steps are executed:

acquiring the fund party matching sequence corresponding to the fund matching strategy;

and sequentially acquiring the matching results fed back by the fund party terminals according to the fund party matching sequence until the fact that the matching result fed back by the fund party terminals is the agreeing matching result is detected for the first time, stopping matching, and acquiring the fund party information corresponding to the fund party terminals, of which the agreeing matching result is detected for the first time, as the fund party matching result corresponding to the target user asset order data.

In this embodiment, when the fund matching policy is adopted for performing the fund matching on all the user asset order data with the first preset type identification result, the order of the fund matching may be referred to the order of the fund matching policy, for example, the order of the fund matching is the order of the class a fund, the class B fund, and the class C fund. For each target user asset order data, the target user asset order data is firstly sent to the fund party terminal of the class A fund party to obtain the feedback matching result. And if the fund party terminal of the A-class fund party feeds back a consent matching result, taking the A-class fund party as the fund party matching result corresponding to the target user asset order data, and not matching with the later B-class fund party and C-class fund party.

However, if the fund side terminal of the class a fund side feeds back a disagreement matching result, the target user asset order data is required to be sent to the fund side terminal of the class B fund side to obtain the feedback matching result, and if the fund side terminal of the class B fund side feeds back an agreement matching result, the class B fund side is used as the fund side matching result corresponding to the target user asset order data, and the matching with the following class C fund side is no longer performed. If the fund side terminal of the B-class fund side feeds back a disagreement matching result, the target user asset order data is required to be sent to the fund side terminal of the C-class fund side to obtain the feedback matching result, and if the fund side terminal of the C-class fund side feeds back the disagreement matching result, the C-class fund side is used as the fund side matching result corresponding to the target user asset order data. If the fund party terminal of the C-type fund party feeds back a disagreement matching result, the result indicates that the target user asset order data is not successfully matched, and a null value valued fund party matching result is obtained. It can be seen that based on the above approach, a fast asset fund match can be made for all user asset order data with a first preset type identification result.

In order to realize order risk identification on the user asset order data and match fund parties after verification of the order risk identification in a server, a matching system with the following architecture can be constructed:

1) The analysis layer can be used for acquiring data such as a historical user asset order data set, a third party user data set and the like stored in the data layer for data analysis to obtain important parameter indexes which are respectively used as output indexes of a subsequent classification prediction model and a decision tree strategy model; for example, the follow-up classification prediction model outputs a prediction result corresponding to the user asset order data, and the decision tree policy model outputs an order risk identification result corresponding to the updated prediction result;

2) The data layer can be used for storing data such as a historical user asset order data set, a third party user data set and the like for other layers to call for data analysis or model training;

3) The model layer is used for storing a two-class prediction model set for primary risk identification of order data, and is generally used for identification;

4) The policy layer is used for storing a policy model set of the decision tree for identifying the secondary risk of the order data;

5) The application layer is used for carrying out asset fund matching on the target user asset order data respectively corresponding to the target order risk identification results identified by the strategy layer;

6) The maintenance layer can be used for maintaining or updating the models or strategies in the model layer and the strategy layer.

Therefore, the embodiment of the method can screen out the target user asset order data with the first preset type identification result to participate in final capital asset matching after two risk evaluations of the two classification prediction model sets and the decision tree strategy model sets are adopted for the user asset order data sets, so that the data utilization rate of the asset order is improved, and the problem that the user asset order data is intercepted by mistake is avoided due to the adoption of secondary risk identification.

FIG. 4 is a schematic block diagram of an asset funds matching device based on order risk identification provided by an embodiment of the invention. As shown in fig. 4, corresponding to the above asset funds matching method based on order risk identification, the present invention further provides an asset funds matching device 100 based on order risk identification. The order risk identification based asset funds matching device 100 comprises means for performing the order risk identification based asset funds matching method described above. Referring to fig. 4, the asset funds matching device 100 based on order risk identification includes: an order data acquisition unit 110, a classification prediction unit 120, a prediction result update unit 130, a decision unit 140, a recognition result screening unit 150, and an asset funds matching unit 160.

The memory storage control unit 110 is configured to respond to the asset funds matching instruction, and acquire a user asset order data set corresponding to the asset funds matching instruction.

In one embodiment, the asset funds matching device 100 based on order risk identification further comprises:

the order feature extraction unit is used for acquiring user features, sponsor features and order features corresponding to the user asset order data in the user asset order data set;

the first input data processing unit is used for forming corresponding user asset order input data by the user features, the sponsor features and the order features corresponding to the user asset order data if the user features, the sponsor features and the order features corresponding to the user asset order data are all determined to be non-null features;

and the second input data processing unit is used for acquiring third party user data of the corresponding user asset order data to extract the third party feature and update the party feature if the party feature corresponding to the user asset order data is determined to be the null feature, and the corresponding user asset order input data is composed of the user feature, the party feature and the order feature of the corresponding user asset order data.

A bi-classification prediction unit 120, configured to obtain a prediction result set corresponding to the user asset order dataset based on a pre-trained bi-classification prediction model set; wherein each user asset order data in the user asset order dataset corresponds to a predicted outcome.

In this embodiment, a set of two-class prediction models formed by a plurality of two-class prediction models that have been trained in advance is stored locally in the server, and when each user asset order data in the user asset order data set is used as an input of one of the two-class prediction models to perform a prediction operation, a prediction result corresponding to each user asset order data can be obtained. Because the first risk assessment based on the user asset order data adopts a two-class model, the method has higher prediction result acquisition efficiency, and simultaneously avoids directly taking the prediction result as a final order risk assessment standard, thereby improving the order utilization rate.

In this embodiment, the set of classified prediction models includes at least a first classified prediction model, a second classified prediction model, a third classified prediction model, and a fourth classified prediction model; the first two-class prediction model is used for carrying out order risk identification on user asset order data with a first order accounting period type; the second classification prediction model is used for carrying out order risk identification on user asset order data with a second order accounting period type; the third classification prediction model is used for carrying out order risk identification on user asset order data with a third order accounting period type; the fourth classification prediction model is used for carrying out order risk identification on user asset order data with a fourth order accounting period type; the account period duration corresponding to the first order account period type is smaller than the account period duration corresponding to the second order account period type, the account period duration corresponding to the second order account period type is smaller than the account period duration corresponding to the third order account period type, and the account period duration corresponding to the third order account period type is smaller than the account period duration corresponding to the fourth order account period type;

Wherein the two-class prediction unit 120 performs the following steps when obtaining a corresponding prediction result for each user asset order data of the user asset order data set based on the two-class prediction model set:

And the prediction result updating unit 130 is configured to obtain the supplemental user features corresponding to each prediction result in the prediction result set to update each prediction result, so as to obtain an updated prediction result set.

In one embodiment, the prediction result updating unit 130 is configured to:

acquiring user asset order data corresponding to each prediction result in the prediction result set;

acquiring a wind control feature set and order supplement features corresponding to each user asset order data to form supplement user features corresponding to each prediction result; the wind control feature sets corresponding to the user asset order data are obtained by inputting the user asset order data into a plurality of wind control models for calculation;

And updating each prediction result by using the corresponding supplementary user characteristics of each prediction result to obtain the updated prediction result set.

The decision unit 140 is configured to obtain an order risk identification result set corresponding to the updated prediction result set based on a pre-trained decision tree policy model set; wherein each updated prediction result in the updated prediction result set corresponds to an order risk identification result.

In this embodiment, the set of decision tree policy models includes a first decision tree policy model corresponding to the first two-class prediction model, a second decision tree policy model corresponding to the second two-class prediction model, a third decision tree policy model corresponding to the third two-class prediction model, and a fourth decision tree policy model corresponding to the fourth two-class prediction model;

wherein, when the decision unit 140 obtains the order risk identification result based on the decision tree policy model set for each updated prediction result of the updated prediction result set, the following steps are executed:

the model evaluation index value acquisition unit is used for acquiring an initial decision tree strategy model set and acquiring model evaluation index values obtained by each initial decision tree strategy model in the initial decision tree strategy model set based on a model test set;

the decision tree screening unit is used for acquiring a corresponding initial decision tree strategy model to form the decision tree strategy model set if the model evaluation index value of the initial decision tree strategy model exceeds the preset evaluation index threshold value.

The recognition result screening unit 150 is configured to obtain the order risk recognition results with the first preset type of recognition results in the order risk recognition result set to form a target order risk recognition result set.

The asset and fund matching unit 160 is configured to obtain target user asset order data corresponding to each target order risk identification result in the target order risk identification result set, and obtain a fund party matching result corresponding to each target user asset order data based on a preset fund matching policy.

In this embodiment, after the target order risk identification result set is obtained in the server, a prestored fund matching policy needs to be obtained for matching the target user asset order data with the fund party, so as to achieve final asset and fund matching.

In an embodiment, the fund matching policy is used for matching the user asset order data according to a preset fund matching order to obtain a fund matching result, and the asset matching unit 160 performs the following steps when determining the fund matching result for each target user asset order data in the target order risk identification result set based on the fund matching policy:

Therefore, the embodiment of the device can screen out the target user asset order data with the first preset type identification result to participate in final capital asset matching after two risk evaluations of the two classification prediction model sets and the decision tree strategy model sets are adopted for the user asset order data sets, so that the data utilization rate of the asset order is improved, and the problem that the user asset order data is intercepted by mistake is avoided due to the adoption of secondary risk identification.

It should be noted that, as those skilled in the art can clearly understand, the specific implementation process of the above-mentioned asset and funds matching device and each unit based on order risk identification may refer to the corresponding description in the foregoing method embodiment, and for convenience and brevity of description, the description is omitted here.

The above-described order risk identification-based asset funds matching device may be implemented in the form of a computer program that may be run on an order risk identification-based computer device as shown in fig. 5.

Referring to fig. 5, fig. 5 is a schematic block diagram of a computer device according to an embodiment of the present invention. The computer equipment integrates any asset fund matching device based on order risk identification provided by the embodiment of the invention.

With reference to fig. 5, the computer device includes a processor 402, a memory, and a network interface 405, which are connected by a system bus 401, wherein the memory may include a storage medium 403 and an internal memory 404.

The storage medium 403 may store an operating system 4031 and a computer program 4032. The computer program 4032 includes program instructions that, when executed, cause the processor 402 to perform an asset funds matching method based on order risk identification.

The processor 402 is used to provide computing and control capabilities to support the operation of the overall computer device.

The internal memory 404 provides an environment for the execution of a computer program 4032 in the storage medium 403, which computer program 4032, when executed by the processor 402, causes the processor 402 to perform the asset funds matching method based on order risk identification described above.

The network interface 405 is used for network communication with other devices. It will be appreciated by those skilled in the art that the structure shown in FIG. 5 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

Wherein the processor 402 is configured to execute the computer program 4032 stored in the memory to implement the asset funds matching method based on order risk identification described above.

It should be appreciated that in embodiments of the present invention, the processor 402 may be a Central processing unit (Central ProcessingUnit, CPU), the processor 402 may also be other general purpose processors, digital signal processors (DigitalSignalProcessor, DSP), application specific integrated circuits (application specific IntegratedCircuit, ASIC), off-the-shelf programmable gate arrays (Field-ProgrammableGateArray, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

Those skilled in the art will appreciate that all or part of the flow in a method embodying the above described embodiments may be accomplished by computer programs instructing the relevant hardware. The computer program comprises program instructions, and the computer program can be stored in a storage medium, which is a computer readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the flow steps of the embodiments of the method described above.

Accordingly, the present invention also provides a computer-readable storage medium. The computer readable storage medium stores a computer program, wherein the computer program includes program instructions. The program instructions, when executed by the processor, cause the processor to perform the above-described asset funds matching method based on order risk identification.

The storage medium may be a U-disk, a removable hard disk, a Read-only memory (ROM), a magnetic disk, or an optical disk, or other various computer-readable storage media that may store program codes.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of each unit is only one logic function division, and there may be another division manner in actual implementation. For example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed.

The steps in the method of the embodiment of the invention can be sequentially adjusted, combined and deleted according to actual needs. The units in the device of the embodiment of the invention can be combined, divided and deleted according to actual needs. In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.

The integrated unit may be stored in a storage medium if implemented in the form of a software functional unit and sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a terminal, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention.

While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims

1. An asset funds matching method based on order risk identification, comprising:

2. The method of claim 1, wherein after the acquiring of the user asset order dataset corresponding to the asset funds matching instruction in response to the asset funds matching instruction, before the acquiring of the prediction result set corresponding to the user asset order dataset based on the pre-trained binary prediction model set, the method further comprises:

3. The method of claim 2, wherein the set of classified prediction models includes at least one classified prediction model, the user asset order data includes at least one order ledger type, different order ledger types have different ledger durations, and the at least one classified prediction model corresponds to the ledger durations of the at least one order ledger type;

4. The method of claim 1, wherein the obtaining the supplemental user features corresponding to each of the predictors in the set of predictors to update each of the predictors to obtain an updated set of predictors comprises:

5. A method according to claim 3, wherein the set of decision tree policy models comprises at least one decision tree policy model, and the at least one decision tree policy model corresponds to the at least one classification prediction model;

6. The method of claim 1, wherein prior to the obtaining, based on the pre-trained set of decision tree policy models, a set of order risk identification results corresponding to the updated set of predicted results, the method further comprises:

7. The method of claim 1, wherein the fund matching policy is configured to match the user asset order data according to a preset fund matching order to obtain a fund matching result;

and executing the steps of acquiring the fund party matching result corresponding to the asset order data of each target user based on the preset fund matching strategy, wherein the fund party matching result is executed for each asset order data of each target user:

8. An asset funds matching device based on order risk identification, comprising:

9. A computer device comprising a memory having a computer program stored thereon and a processor that when executed implements the order risk identification-based asset funds matching method of any of claims 1-7.

10. A computer readable storage medium, characterized in that it stores a computer program comprising program instructions which, when executed by a processor, implement the order risk identification based asset fund matching method as defined in any one of claims 1-7.