Information Processing Method, Device, and Computer Equipment for Target Objects

Abstract

The present disclosure relates to an information processing method, apparatus, and computer device for a target object. The method includes: receiving a cooperation request initiated by the target object; obtaining a score of the risk factors of the target object and a total approval score; based on the score of the risk factors and the total approval score, receiving a first approval result for the target object; labeling the target object for which the first approval result is disapproval, and storing the identity information, the score of the risk factors, the total approval score, the first approval result, and the label of the target object for which the first approval result is disapproval in a sample database; constructing a machine learning approval model according to the sample data. The present disclosure can not only reduce the dependence on manual work and improve the approval efficiency, but also obtain the scores of various risk factors of the target object, which is helpful for subsequent verification of the machine learning approval model.

Description

Technical Field

[0001] The present disclosure relates to the field of information processing technologies, and in particular, to an information processing method, apparatus, computer device for a target object. Background Art

[0002] Informatization is a major trend in the development of the current era, and information industrialization has also become an historical trend. Informatization can provide an important guarantee for enterprises to guard against uncertainties and help managers enhance their decision-making capabilities. Especially in the banking field of financial transactions, some businesses are actively shifting from offline to online. In terms of the evaluation of open bank partners, although banks have mature risk control processes for the approval of traditional enterprise business, the approval rules can be interpreted item by item through corresponding regulatory measures and implemented; however, it is difficult to apply to the actual evaluation methods of open bank partners. Open bank partners are the evolution of user roles under the innovation of traditional business models, but the driving force for this evolution is the reshaping of user experience by open banks and the promotion of demand innovation in the business journey.

[0003] Therefore, with the expansion of business scenarios, transaction processes and partners have become complex and diverse, and open bank risk control faces new challenges. Traditional enterprise approval methods are no longer applicable to open banks, and there is an urgent need to propose a new evaluation method for partner information processing. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide an information processing method, apparatus, computer device, storage medium and computer program product for a target object.

[0005] In a first aspect, the present disclosure provides an information processing method for a target object. The method includes:

[0006] Receiving a cooperation request initiated by a target object, and obtaining the identity information of the target object based on the cooperation request;

[0007] Using a pre-constructed expert scoring model to approve the target object, and obtaining the scores of the risk factors of the target object and the total approval score;

[0008] Based on the scores of the risk factors and the total approval score, receiving a first approval result for the target object; the first approval result includes approval and disapproval;

[0009] Labeling the target object with a disapproval in the first approval result, and storing the identity information, scores of the risk factors, total approval score, first approval result of the target object, and the label of the target object with a disapproval in the first approval result in a sample database;

[0010] Obtain sample data from the sample database, construct a machine learning approval model based on the sample data, and use the machine learning approval model to perform approval processing on the target object. The machine learning approval model is used to output the second approval result of the target object and the score of the risk factor of the target object.

[0011] In one embodiment, before using the pre-constructed expert scoring model to approve the target object and obtain the score of the risk factor of the target object and the total approval score, it further includes:

[0012] According to the identity information of the target object, perform mandatory rule screening on the target object.

[0013] In one embodiment, the construction steps of the expert scoring model include:

[0014] Receive the risk factors set for the target object, the algorithm rules set for the risk factors, and the weights set for the risk factors;

[0015] Establish the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is used to output the score of the risk factor of the target object and the total approval score according to the identity information of the target object.

[0016] In one embodiment, the marking of the target object with the first approval result of disapproval includes:

[0017] Mark the target object with disapproval. Select the label of the reason for the disapproval of the target object on the marking list, sort the labels according to the preset rules, and store the labels and label order in the sample database;

[0018] Optimize and update the expert scoring model according to the labels and label order of the target object.

[0019] In one embodiment, the construction steps of the machine learning approval model include:

[0020] Obtain the sample data of the first approval result of the target object from the sample database. The sample data includes the first approval result of the target object, the score of the risk factor of the target object, and the marking of the target object with the first approval result of disapproval;

[0021] A risk sub-model is established based on the identity information of the target object and the scores of risk factors in the sample data, and an approval result sub-model is established based on the identity information of the target object and the first approval result in the sample. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-model corresponds to the risk factor one by one.

[0022] In one embodiment, the establishing a risk sub-model based on the identity information of the target object and the scores of risk factors in the sample data, and establishing an approval result sub-model based on the identity information of the target object and the first approval result in the sample, where the risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-model corresponding to the risk factor one by one includes:

[0023] Extract the input model feature variables according to the identity information of the target object;

[0024] Based on the linear regression algorithm, calculate the input model feature variables to obtain the probability ratio of approval and non-approval;

[0025] Convert the probability ratio into the scores of the risk factors and the total score of the target object;

[0026] Conduct an effectiveness test on the risk sub-model and the approval result sub-model. If the test passes, establish the machine learning approval model.

[0027] In one embodiment, the extracting the input model feature variables according to the identity information of the target object includes:

[0028] Extract feature variables according to the identity information of the target object to form a feature wide table, and conduct a preliminary screening on the feature wide table to form a candidate feature pool;

[0029] Conduct a secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables;

[0030] Conduct chi-square binning on the effective feature variables, adjust the coordinate point threshold of the binning, and determine the input model feature variables according to the evidence weight of each bin.

[0031] In a second aspect, the present disclosure also provides an information processing device for a target object. The device includes:

[0032] A cooperation request module, configured to receive a cooperation request initiated by a target object, and obtain the identity information of the target object based on the cooperation request;

[0033] An expert scoring module, configured to approve the target object by using a pre-constructed expert scoring model, and obtain the scores of the risk factors and the total approval score of the target object;

[0034] An expert approval module, configured to receive a first approval result for the target object based on the score of the risk factor and the total approval score; the first approval result includes approval and disapproval.

[0035] A sample module, configured to store the identity information of the target object, the score of the risk factor, the total approval score, and the first approval result in a sample database.

[0036] A model approval module, configured to obtain sample data from the sample database, construct a machine learning approval model based on the sample data, and use the machine learning approval model to approve the target object. The machine learning approval model is configured to output a second approval result for the target object and the score of the risk factor of the target object.

[0037] In one embodiment, the apparatus further includes:

[0038] A forced screening module, configured to perform forced rule screening on the target object according to the identity information of the target object.

[0039] In one embodiment, the expert scoring module includes:

[0040] An expert setting unit, configured to receive risk factors set for the target object, algorithm rules set for the risk factors, and weights set for the risk factors.

[0041] A score calculation unit, configured to establish the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is configured to output the score of the risk factor and the total approval score of the target object according to the identity information of the target object.

[0042] In one embodiment, the sample module further includes:

[0043] A labeling unit, configured to label the target object with disapproved approval, select labels for the reasons for the disapproved approval of the target object on a labeling list, sort the labels according to a preset rule, and store the labels and the label order in the sample database.

[0044] An optimization unit, configured to optimize and update the expert scoring model according to the labels and the label order of the target object.

[0045] In one embodiment, the model approval module includes:

[0046] A sample extraction module, configured to obtain sample data of the first approval result of a target object from the sample database, where the sample data includes the first approval result of the target object, the scores of the risk factors of the target object, and the annotations of the target objects with the first approval result being disapproved;

[0047] A model establishment module, configured to establish a risk sub-model based on the identity information of the target object and the scores of the risk factors in the sample data, and establish an approval result sub-model based on the identity information of the target object and the first approval result in the sample. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-models correspond to the risk factors one by one.

[0048] In one embodiment, the model establishment module includes:

[0049] A sample processing unit, configured to extract input model feature variables according to the identity information of the target object;

[0050] A regression algorithm unit, configured to calculate the input model feature variables based on a linear regression algorithm to obtain the probability ratio of approval and disapproval;

[0051] A score output unit, configured to convert the probability ratio into the scores of the risk factors and the total score of the target object;

[0052] A verification unit, configured to perform validity verification on the risk sub-model and the approval result sub-model, and establish the machine learning approval model when the verification passes.

[0053] In one embodiment, the sample processing unit includes:

[0054] A preliminary screening subunit, configured to extract feature variables according to the identity information of the target object to form a feature wide table, and perform preliminary screening on the feature wide table to form a candidate feature pool;

[0055] A secondary screening subunit, configured to perform secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables;

[0056] An input model feature variable unit, configured to perform chi-square binning on the effective feature variables, adjust the coordinate point threshold of the binning, and determine the input model feature variables according to the evidence weight of each bin.

[0057] In a third aspect, the present disclosure also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the above information processing method for target objects are implemented.

[0058] Fourthly, the present disclosure also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the steps of the above-mentioned information processing method for a target object are implemented.

[0059] Fifthly, the present disclosure also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned information processing method for a target object are implemented.

[0060] The above-mentioned information processing method, device, computer device, storage medium, and computer program product for a target object have at least the following beneficial effects:

[0061] The present disclosure combines an expert scoring model and a machine learning approval model, realizing the transition from the cold start stage to the model intelligent decision-making stage. The expert scoring model provides sample data for the construction of the machine learning approval model, and constructs a sample database while completing the approval service. The machine learning approval model can approve a target object based on the identity information of the target object, which can not only reduce the dependence on manual work to obtain the approval result and improve the approval efficiency, but also obtain the scores of various risk factors of the target object. The scores of the risk factors are helpful for subsequent verification, optimization iteration, and improvement of the accuracy of the machine learning approval model. Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or in the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Figure 1 It is an application environment diagram of the information processing method for a target object in an embodiment;

[0064] Figure 2 It is a flowchart of the information processing method for a target object in an embodiment;

[0065] Figure 3 It is another flowchart of the information processing method for a target object in an embodiment

[0066] Figure 4 It is a flowchart of the steps for constructing an expert scoring model in an embodiment;

[0067] Figure 5 It is a flowchart of the annotation steps in an embodiment;

[0068] Figure 6 is a flowchart of an information processing method for a target object in an embodiment;

[0069] Figure 7 is a schematic flowchart of the steps for constructing a machine learning approval model in an embodiment;

[0070] Figure 8 is another schematic flowchart of the steps for constructing a machine learning approval model in an embodiment;

[0071] Figure 9 is a schematic flowchart of the steps for extracting input model feature variables in an embodiment;

[0072] Figure 10 is a flowchart of the steps for constructing a machine learning approval model in an embodiment;

[0073] Figure 11 is a block diagram of a structure of an information processing device for a target object in an embodiment;

[0074] Figure 12 is another block diagram of a structure of an information processing device for a target object in an embodiment;

[0075] Figure 13 is a block diagram of a structure of an expert scoring module in an embodiment;

[0076] Figure 14 is a block diagram of a structure of a sample module in an embodiment;

[0077] Figure 15 is a block diagram of a structure of a model approval module in an embodiment;

[0078] Figure 16 is a block diagram of a structure of a model establishment module in an embodiment;

[0079] Figure 17 is a block diagram of a structure of a sample processing unit in an embodiment;

[0080] Figure 18 is an internal block diagram of a computer device in an embodiment. Detailed implementation manners

[0081] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0082] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. The terms used in the description of this disclosure herein are for the purpose of describing specific embodiments only and are not intended to limit this disclosure.

[0083] It should be noted that the terms "first", "second", etc. in the description and claims of this disclosure and the above-mentioned drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of this disclosure as detailed in the appended claims. The term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, such that a process, method, product or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, product or apparatus. Without further limitation, there is no exclusion of the presence of additional identical or equivalent elements in the process, method, product or apparatus comprising the said elements. For example, if the words first, second, etc. are used to denote names, they do not denote any particular order.

[0084] As used herein, the singular forms "a", "an" and "the" may also include the plural forms unless the context clearly dictates otherwise. It should also be understood that the terms "comprising / including" or "having" etc. specify the presence of the stated features, wholes, steps, operations, components, parts or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof. At the same time, in this specification, the term "and / or" includes any and all combinations of the related listed items.

[0085] The information processing method for a target object provided by the embodiments of this application can be applied to an application environment as Figure 1 shown. Among them, the client 102 communicates with the server 104 through a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104 or placed in the cloud or on other network servers. Among them, the server 104 can be implemented by an independent server or a server cluster composed of multiple servers. For example, the client 102 can be a terminal or a server device for providing a service for an enterprise or individual that submits a cooperation application to a bank to initiate a cooperation request to the server 104.

[0086] In some embodiments of the present disclosure, as Figure 2 shown, an information processing method for a target object is provided. Taking the server in Figure 1 as an example for illustration, the method includes the following steps:

[0087] Step S10: Receive a cooperation request initiated by the target object, and obtain the identity information of the target object based on the cooperation request.

[0088] Specifically, the target object may initiate a cooperation request through an online device. In this embodiment, after receiving the cooperation request from the target object, the identity information of the target object can be obtained. Here, the identity information may include the industrial and commercial registration information of the target object, personnel flow conditions, revenue details, whether it has received administrative penalties, market share, etc.

[0089] Step S30: Use a pre-constructed expert scoring model to approve the target object, and obtain the scores of the risk factors of the target object and the total approval score.

[0090] Specifically, the expert scoring model generally refers to making a quantitative evaluation in the form of scoring based on quantitative and qualitative analysis, and its results have mathematical statistical characteristics. The greatest advantage of the expert scoring model is that it can make a quantitative estimate in the absence of sufficient statistical data and original materials. In this embodiment, in the cold start stage, the risk factors of the target object are scored respectively through the expert scoring model, and the total approval score is obtained.

[0091] Among them, the risk factor may refer to the enterprise risk factor of the target object, the conditions that can promote or cause the occurrence of a risk event, and the conditions that cause the loss to increase and expand when the risk event occurs. The risk factor is a potential factor for the occurrence of a risk event and an indirect and internal cause of the loss. For example, the risk factors of the target object may include the management ability, technical ability, risk control and security ability, and market competitiveness of the target object. The expert scoring model can conduct a detailed analysis and comparison of each risk factor by a panel of authoritative experts in this field, and finally obtain the score of each risk factor and the total approval score by synthesizing the scores of each risk factor.

[0092] Step S50: Based on the scores of the risk factors and the total approval score, receive the first approval result for the target object; the first approval result includes approval and non-approval.

[0093] Specifically, after the expert scoring model outputs the scores of the risk factors and the total approval score, the expert panel gives the first approval result of approval or non-approval. The first approval result for the target object is received through an input device.

[0094] Step S70: Label the target objects with a first approval result of disapproval, and store the identity information of the target objects, the scores of risk factors, the total approval score, the first approval result, and the labels of the target objects with a first approval result of disapproval in the sample database.

[0095] Specifically, when storing the first approval result, the scores of risk factors, and the total approval score output by the expert scoring model for the target objects in the sample database, at the same time, regard the target objects with a disapproved first approval result as bad samples, label the bad samples, and store the labeling results in the sample database.

[0096] Step S90: Obtain sample data from the sample database, construct a machine learning approval model based on the sample data, and use the machine learning approval model to perform approval processing on the target objects. The machine learning approval model is used to output the second approval result of the target objects and the scores of the risk factors of the target objects.

[0097] Specifically, based on the constructed sample database, a machine learning approval model is established. Use the first approval result in the sample database and the process data for obtaining the first approval result (the scores of risk factors and the total approval score output by the expert scoring model, the labels of bad samples) as the learning samples of the machine learning approval model, and construct the machine learning approval model through machine learning. After the machine learning approval model learns enough sample data, a mature machine learning model is formed. The machine learning approval model can output the second approval result of the target object and the scores of the risk factors of the target object after inputting the identity information of a new target object that is non-sample data.

[0098] In the above information processing method for target objects, the expert scoring model and the machine learning approval model are combined to realize the transition from the cold start stage to the model intelligent decision-making stage. The expert scoring model provides sample data for the construction of the machine learning approval model, and constructs the sample database while completing the approval business. The machine learning approval model can approve the target objects based on the identity information of the target objects, which can not only reduce the dependence on manual work to obtain the approval result and improve the approval efficiency, but also obtain the scores of each risk factor of the target objects. The scores of the risk factors are helpful for subsequent verification, optimization and iteration of the machine learning approval model, and improve the accuracy of the machine learning approval model.

[0099] In some embodiments of the present disclosure, as Figure 3 shown, before the above step S30, it further includes:

[0100] Step S20: Perform forced rule screening on the target objects according to the identity information of the target objects.

[0101] Specifically, in order to reduce subsequent processing of the identity information of the target object, the target object requesting cooperation can be preliminarily screened by setting mandatory rules. The mandatory rules can be set to obtain a yes or no output result without algorithm calculation or only through simple algorithm calculation, and the target objects with a yes output result pass the screening. One or more mandatory rules can be set. For example, a mandatory rule can be set that when the market share is less than the set threshold, the output is no.

[0102] In some embodiments of the present disclosure, as Figure 4 shown, the steps for constructing the expert scoring model include:

[0103] Step A10: Receive the risk factors set for the target object, the algorithm rules set for the risk factors, and the weights set for the risk factors.

[0104] Specifically, domain experts of the target object are selected offline or online to select the risk factors of the target object. When selecting risk factors, the risk factors required for the subsequent constructed machine learning approval model can be evaluated. Design a long list of risk factors for the target object, and after recovering the long list of risk factors, statistically analyze the evaluation results. The long list of risk factors can reflect the enterprise risk impact of the target object. After integrating the opinions of each expert, the long list of risk factors mainly includes the management ability, technical ability, risk control and security ability, and market competitiveness of the target object. After determining the long list of risk factors, initiate a new round of risk factor evaluation of the target object to gradually converge the opinions of the experts and finally form a short list of risk factors, that is, the finally set risk factors. After the experts set the risk factors, they can set algorithm rules and the weights corresponding to each risk factor according to the degree of importance or other bases. This method can receive the risk factors set by the experts, the algorithm rules set for the risk factors, and the weights set for the risk factors through offline or online input.

[0105] Step A20: Establish the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is used to output the scores of the risk factors of the target object and the total approval score according to the identity information of the target object.

[0106] Specifically, according to the received risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors, construct an expert scoring model. The expert scoring model can calculate the scores of each evaluation factor according to the identity information of the target object, and perform weighting according to the scores of the evaluation factors to obtain the total approval score.

[0107] In this embodiment, by consulting expert experience, the cooperation requests of the target objects are analyzed and summarized, and an expert scoring model is established. Especially for the situation where there is a lack of data on the target objects and quantitative analysis cannot rely on a large amount of data, this method is made more objective and accurate.

[0108] In some embodiments of the present disclosure, as Figure 5 shown, the above step S70 includes:

[0109] Step S72: Label the target objects that fail the approval, select the labels for the reasons why the target objects fail the approval on the labeling list, sort the labels according to a preset rule, and store the labels and the label order in the sample database.

[0110] Specifically, when storing the first approval result output by the expert scoring model for the target object, the scores of the risk factors, and the total approval score in the sample database, at the same time, store the labels of the target objects with the first approval result being not approved in the sample database. When labeling the target objects with the first approval result being not approved, introduce a labeling list. For example, the labeling list can consist of major categories (security / technology / funds / credit / market / others), and sub-categories under the corresponding major categories (poor information security / loss of core personnel / substantial decline in quarterly earnings / subjected to administrative penalties / shrinkage of market share). For the target objects with the first approval result being not approved, check the eligible labels under each category in the order of importance of the reasons for non-approval as further explanations, and store these labeling results in the sample database.

[0111] Step S74: Optimize and update the expert scoring model according to the labels and the label order of the target object.

[0112] Specifically, through annotation, tags can be added to the target objects, and then the expert scoring model can be optimized and updated. For example, target objects with the first approval result of disapproval can be periodically selected, their annotation content can be obtained, and compared with the scores and weights of each risk factor of the expert scoring model to screen out the deviation between the expert scoring model and the annotation. Through a priori and posteriori, the deviation between the expert scoring model and the actual business can be found in a timely manner. For example, whether the weight ranking of each risk factor is in good consistency with the ranking of the annotation content; whether the risk factor score of the target object matches the annotation content. First, find the deviation, and then explore whether the business logic has been updated or the deviation of the expert scoring model itself, so as to optimize and iterate the expert scoring model. Changes in the business scenario and modifications to the pre-set rules will cause survivor bias. The annotation list needs to be continuously updated according to the current business knowledge and also be traceable to historical decisions. Store the annotation rules with the annotation rule ID (Identity document) and time as the primary key to ensure the data traceability during offline analysis and batch processing.

[0113] Meanwhile, the data source can also be verified through annotation. Summarize the currently connected external data sources and check whether the annotation content shows good correlation with the data source content. For example, whether there are contradictory situations between the target objects rejected due to small capital scale and their corresponding public opinion data and industrial and commercial data.

[0114] In addition, this method can also monitor the annotation, regularly count the annotation content of the target objects, including the major categories and minor categories under the annotation list, form a report analysis, and track the changes of the target objects. Automatically monitor the annotation content and the scores of each risk factor of the expert scoring model to form an intelligent report, which helps to optimize and update the expert scoring model.

[0115] In this embodiment, by introducing annotation in the approval process, the target objects with the first approval result of disapproval output by the expert scoring model are annotated, providing structured analysis data for the portrait of the target objects, including operations such as classifying, sorting, editing, correcting errors, marking, and commenting on the annotation data such as the external risk data of the target objects, adding tags to the target objects, facilitating the stage verification of the expert scoring model, and also meeting the requirements of subsequent machine learning training based on data samples.

[0116] In some embodiments of the present disclosure, as Figure 6 and Figure 7 shown, the construction steps of the machine learning approval model include:

[0117] Step B10: Obtain sample data of the first approval result of the target object from the sample database, where the sample data includes the first approval result of the target object, the scores of the risk factors of the target object, and the annotations of the target objects with a non-passed first approval result.

[0118] Specifically, in the early stage of constructing the machine learning approval model, sufficient sample data needs to be learned. The sample data can be obtained from a pre-constructed sample database. The sample data can include the first approval result of the target object, the scores of the risk factors of the target object, and the annotations of the target objects with a non-passed first approval result.

[0119] Step B20: Establish a risk sub-model based on the identity information and the scores of the risk factors of the target object in the sample data, and establish an approval result sub-model based on the identity information and the first approval result of the target object in the sample. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-models correspond to the risk factors one by one.

[0120] Specifically, the machine learning approval model can include several risk sub-models and an approval result sub-model. The risk sub-models can be directly established with the risk factors set by the expert scoring model, or combined with the annotations to analyze and summarize the set of risk factors that are most desired to capture. For example, if there are 3 risk factors in the risk factor set, then for these 3 risk factors, 3 risk sub-models are established in combination with the sample data and the annotations, and the risk sub-models correspond to the risk factors one by one. The risk sub-models are used to output the scores of the corresponding risk factors. An approval result sub-model is established based on the first approval result output by the expert scoring model, and the approval result sub-model is used to output the second approval result.

[0121] The approval result sub-model can also replace the expert scoring model in the cold start stage to approve the historical sample data, obtain the second approval result and the total approval score of the sample data, select the confidence interval of the total approval score output by the approval result sub-model according to the set rules, and determine it as the score interval for the second approval result being not passed, and the operation of not passing the approval can be directly performed by the approval result sub-model.

[0122] This embodiment helps to provide structured analysis data for the portrait of the target object and display the risks of the target object in multiple dimensions by establishing risk sub-models for each risk factor and an approval result sub-model based on whether the approval is passed.

[0123] In some embodiments of the present disclosure, as Figure 8 shown, the above step B20 includes:

[0124] Step B22: Extract the input model feature variables according to the identity information of the target object.

[0125] Specifically, summarize the sample data in the sample database, extract the input model variables from the identity information of the target object, and use the input model variables as independent variables.

[0126] Step B24: Calculate the input model feature variables based on the linear regression algorithm to obtain the probability ratio of approval passing and not passing.

[0127] Specifically, perform algorithm calculation on the selected input model feature variables to obtain the probability ratio of approval passing and not passing for the current target object. In this embodiment, the input model feature variables are calculated based on the linear regression algorithm, and other data analysis algorithms can also be used.

[0128] Step B26: Convert the probability ratio into the score of the risk factor of the target object and the total approval score.

[0129] Specifically, convert the probability ratio obtained in step B24 into a scoring system to obtain the total approval score of the current target object. At the same time, calculate the score of the risk factor of the current target object according to the input model feature variables to complete the quantification of the risk factor.

[0130] Step B28: Conduct validity tests on the risk sub-model and the approval result sub-model. In the case of passing the tests, establish the machine learning approval model.

[0131] Specifically, after establishing the risk sub-model and the approval result sub-model, the risk sub-model and the approval result sub-model can be subjected to validity tests through the ROC (receiver operating characteristic) curve, KS (Kolmogorov-Smirnov) statistic, or KS (Kolmogorov-Smirnov) matrix. If the tests are passed, the risk sub-model and the approval result sub-model jointly constitute the machine learning approval model of this method.

[0132] If the test does not meet the expectations, the data at that time can be traced back through the annotation rule ID and the identity information of the target object for annotation optimization, and a batch of labels can be updated offline. It is also possible to optimize the mandatory rules for further screening in the mandatory rule screening stage of step S20 above. It is also possible to perform data processing on the input model feature variables, such as adjusting the binning method of the WOE (Weight of Evidence) of the input model feature variables, optimizing the algorithm model, for example, using decision tree binning and maximum IV (Information Value) binning to perform more accurate discretization processing on continuous data; it is also possible to perform multicollinearity and feature significance screening on the input model feature variables.

[0133] In this embodiment, the input feature variables are selected based on the identity information of the target object, a risk sub-model and an approval result sub-model are constructed based on the linear regression algorithm, and the effectiveness of the risk sub-model and the approval result sub-model is verified, which helps to enhance the generalization of the finally constructed machine learning approval model.

[0134] In some embodiments of the present disclosure, as Figure 9 shown, the above step B22 includes:

[0135] Step B222: Extract feature variables according to the identity information of the target object to form a feature wide table, and perform a preliminary screening on the feature wide table to form a candidate feature pool.

[0136] Specifically, summarize all available data sources in the sample data to form a feature wide table. A round of rough screening can be performed according to the coverage rate of the data source and the feature variance, and invalid features with low coverage rate and variance close to 0 are removed to form a candidate feature pool.

[0137] Step B224: Perform a secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables.

[0138] Specifically, perform a secondary screening on the feature variables in the candidate feature pool. For example, first perform equal-frequency binning on the feature variables in the candidate feature pool, use IV (Information Value, feature information value) to screen effective feature variables, calculate the Peterson coefficient between the effective feature variables, and screen out features with high collinearity to obtain effective feature variables.

[0139] Step B226: Perform chi-square binning on the effective feature variables, adjust the coordinate point threshold of the binning, and determine the input feature variables according to the weight of evidence of each bin.

[0140] Specifically, more refined chi-square binning can be performed on the effective feature variables. Through monotonicity analysis, adjust the coordinate point threshold of the binning, and recalculate the WOE (Weight of Evidence, weight of evidence) of each bin to ensure the interpretability, effect, and stability of the input feature variables.

[0141] In this embodiment, by extracting and analyzing the identity information of the target object, feature variables are obtained, and the feature variables are screened and optimized multiple times, so that the final input feature variables have interpretability and stability, and at the same time help to improve the accuracy of the machine learning approval model.

[0142] In some embodiments of the present disclosure, as Figure 10As shown in the figure, this method includes a cold start phase and an intelligent decision-making flow phase. Among them, in the cold start phase, a cooperation request initiated by a target object is received. First, the target object is screened by mandatory rules. The target object that passes the screening is input into an expert scoring model. The expert scoring model outputs the risk factor score and the total approval score of the target object, receives the first approval result given by the expert, and stores the first approval result in the database.

[0143] By extracting the first approval result output by the expert scoring model, the target objects that fail the approval are labeled. The practice extraction service is used to extract the sample target objects with the first approval result being disapproved, and verify the deviation between their labeled content and the expert scoring model.

[0144] In the intelligent decision-making flow phase, sample features are extracted from the sample database, and equal-frequency pre-binning processing and screening are performed on the sample features to construct an effective feature pool. More refined chi-square binning is performed on the effective feature variables. Through monotonicity analysis, the coordinate point thresholds of the binning are adjusted, and the WOE (Weight of Evidence) of each bin is recalculated to ensure the interpretability, effectiveness, and stability of the feature variables included in the model. The selected feature variables included in the model are trained through logistic regression to construct a machine learning approval model.

[0145] Based on the business scenario of open banking, this embodiment involves the introduction of expert experience in the early stage, the establishment of a labeling system in the middle stage, and the development of a machine learning model in the later stage from the cold start phase established in the initial stage to the intelligent decision-making flow phase in the middle stage. This method can not only accumulate sample labels through the upgrade of business cognition but also access algorithms to accelerate the efficiency of manual review.

[0146] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0147] Based on the same inventive concept, embodiments of the present disclosure further provide an information processing apparatus for a target object for implementing the information processing method for a target object involved above. The implementation solutions provided by this apparatus for solving problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the information processing apparatus for a target object provided below can refer to the limitations on the information processing method for a target object in the foregoing text, and will not be elaborated herein.

[0148] The apparatus may include a system (including a distributed system), software (application), module, component, server, client, etc. that uses the method described in the embodiments of this specification, and an apparatus that combines necessary implementation hardware. Based on the same innovative concept, the apparatus in one or more embodiments provided by the embodiments of the present disclosure is as described in the following embodiments. Since the implementation solutions of the apparatus for solving problems are similar to those of the method, the implementation of the specific apparatus in the embodiments of this specification can refer to the implementation of the foregoing method, and the repeated parts will not be elaborated. As used hereinafter, the term "unit" or "module" may be a combination of software and / or hardware that can implement a predetermined function. Although the apparatuses described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0149] In some embodiments of the present disclosure, as Figure 11 shown, an information processing apparatus for a target object is provided. The apparatus may be the aforementioned server, or a module, component, device, unit, etc. integrated in the terminal. The apparatus Z00 may include:

[0150] A cooperation request module Z10, configured to receive a cooperation request initiated by a target object, and obtain the identity information of the target object based on the cooperation request;

[0151] An expert scoring module Z20, configured to approve the target object by using a pre-constructed expert scoring model, and obtain a score of the risk factor of the target object and a total approval score;

[0152] An expert approval module Z30, configured to receive a first approval result for the target object based on the score of the risk factor and the total approval score; the first approval result includes approval and disapproval;

[0153] A sample module Z40, configured to label the target object for which the first approval result is disapproval, and store the identity information of the target object, the score of the risk factor, the total approval score, the first approval result, and the label of the target object for which the first approval result is disapproval in a sample database;

[0154] The model approval module Z50 is used to obtain sample data from the sample database, construct a machine learning approval model according to the sample data, and use the machine learning approval model to approve the target object. The machine learning approval model is used to output the second approval result of the target object and the score of the risk factor of the target object.

[0155] In some embodiments of the present disclosure, as Figure 12 shown, the device Z00 further includes:

[0156] The forced screening module Z60 is used to perform forced rule screening on the target object according to the identity information of the target object.

[0157] In some embodiments of the present disclosure, as Figure 13 shown, the expert scoring module Z20 includes:

[0158] The expert setting unit Z22 is used to receive the risk factors set for the target object, the algorithm rules set for the risk factors, and the weights set for the risk factors;

[0159] The scoring calculation unit Z24 is used to establish the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is used to output the score of the risk factor of the target object and the total approval score according to the identity information of the target object.

[0160] In some embodiments of the present disclosure, as Figure 14 shown, the sample module Z40 includes:

[0161] The annotation unit Z42 is used to annotate the target object that fails the approval, select the label of the reason for the failure of the target object's approval on the annotation list, sort the labels according to the preset rules, and store the labels and the label order in the sample database;

[0162] The optimization unit Z44 is used to optimize and update the expert scoring model according to the labels and the label order of the target object.

[0163] In some embodiments of the present disclosure, as Figure 15 shown, the model approval module Z50 includes:

[0164] The sample extraction module Z52 is used to obtain the sample data of the first approval result of the target object from the sample database. The sample data includes the first approval result of the target object, the score of the risk factor of the target object, and the annotation of the target object whose first approval result is disapproval.

[0165] A model building module Z54 is configured to build a risk sub-model based on the identity information of the target object and the scores of risk factors in the sample data, and build an approval result sub-model based on the identity information of the target object and the first approval result in the sample. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-model corresponds to the risk factor one by one.

[0166] In some embodiments of the present disclosure, as Figure 16 shown, the model building module Z54 includes:

[0167] A sample processing unit Z542 is configured to extract input model feature variables according to the identity information of the target object;

[0168] A regression algorithm unit Z544 is configured to calculate the input model feature variables based on a linear regression algorithm to obtain the probability ratio of approval and non-approval;

[0169] A score output unit Z546 is configured to convert the probability ratio into the scores of the risk factors of the target object and the total score;

[0170] A verification unit Z548 is configured to perform validity verification on the risk sub-model and the approval result sub-model, and build the machine learning approval model when the verification passes.

[0171] In some embodiments of the present disclosure, as Figure 17 shown, the sample processing unit Z542 includes:

[0172] A preliminary screening subunit Z5422 is configured to extract feature variables according to the identity information of the target object to form a feature wide table, and perform preliminary screening on the feature wide table to form a candidate feature pool;

[0173] A secondary screening subunit Z5424 is configured to perform secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables;

[0174] An input model feature variable unit Z5426 is configured to perform chi-square binning on the effective feature variables, adjust the coordinate point threshold of the binning, and determine the input model feature variables according to the evidence weight of each bin.

[0175] Each module in the above information processing device for a target object can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules. It should be noted that the division of modules in the embodiments of the present disclosure is illustrative, only a logical function division, and there may be other division methods in actual implementation.

[0176] Based on the description of the embodiments of the foregoing information processing method for a target object, in another embodiment provided by the present disclosure, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as Figure 18 shown. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements an information processing method for a target object.

[0177] Those skilled in the art can understand that the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0178] Based on the description of the embodiments of the foregoing information processing method for a target object, in another embodiment provided by the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps in each of the above method embodiments.

[0179] Based on the description of the embodiments of the foregoing information processing method for a target object, in another embodiment provided by the present disclosure, a computer program product is provided, including a computer program. When the computer program is executed by a processor, it implements the steps in each of the above method embodiments.

[0180] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present disclosure are all information and data that have been authorized by the user or fully authorized by all parties.

[0181] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0182] In the description of this specification, the descriptions referring to terms such as "some embodiments", "other embodiments", "ideal embodiments", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example.

[0183] It can be understood that the various embodiments of the above methods in this specification are all described in a progressive manner. The same / similar parts between the various embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. For the relevant parts, refer to the descriptions of other method embodiments.

[0184] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features of the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0185] The above-described embodiments only represent several implementation manners of the present disclosure. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present disclosure, several deformations and improvements can still be made, and these all belong to the protection scope of the present disclosure. Therefore, the protection scope of the patent of the present disclosure should be subject to the appended claims.

Claims

1. An information processing method for a target object, characterized in that The method includes: Receiving a cooperation request initiated by a target object, and obtaining the identity information of the target object based on the cooperation request; Using a pre-constructed expert scoring model to approve the target object, and obtaining the scores of the risk factors of the target object and the total approval score; Based on the scores of the risk factors and the total approval score, receiving a first approval result for the target object, where the first approval result includes approval and non-approval; Labeling the target object with a non-approval first approval result, and storing the identity information of the target object, the scores of the risk factors, the total approval score, the first approval result, and the label of the target object with a non-approval first approval result in a sample database; Obtaining sample data from the sample database, constructing a machine learning approval model according to the sample data, and using the machine learning approval model to approve the target object. The machine learning approval model is used to output a second approval result of the target object and the scores of the risk factors of the target object; Among them, the construction steps of the machine learning approval model include: Obtaining sample data of the first approval result of the target object from the sample database, where the sample data includes the first approval result of the target object, the scores of the risk factors of the target object, and the label of the target object with a non-approval first approval result; Establishing a risk sub-model according to the identity information and the scores of the risk factors of the target object in the sample data, and establishing an approval result sub-model according to the identity information and the first approval result of the target object in the sample data. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-model corresponds to each risk factor one by one.

2. The method according to claim 1, wherein Before using the pre-constructed expert scoring model to approve the target object and obtaining the scores of the risk factors of the target object and the total approval score, it further includes: Performing a mandatory rule screening on the target object according to the identity information of the target object.

3. The method according to claim 1, wherein The construction steps of the expert scoring model include: Receiving the risk factors set for the target object, the algorithm rules set for the risk factors, and the weights set for the risk factors; Establishing the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is used to output the scores of the risk factors of the target object and the total approval score according to the identity information of the target object.

4. The method according to claim 3, characterized in that, The labeling of the target object with a non-approval first approval result includes: Labeling the target object with a non-approval, selecting the label of the reason for the non-approval of the target object on the labeling list, sorting the labels according to a preset rule, and storing the labels and the label order in the sample database; Optimizing and updating the expert scoring model according to the labels and the label order of the target object.

5. The method according to claim 1, characterized in that A risk sub-model is established based on the identity information of the target object and the scores of risk factors in the sample data, and an approval result sub-model is established based on the identity information of the target object and the first approval result in the sample data. The risk sub-model and the approval result sub-model constitute the machine learning approval model; The risk sub-model corresponds one-to-one with the risk factors and includes: Extracting input feature variables according to the identity information of the target object; Calculating the input feature variables based on a linear regression algorithm to obtain the probability ratio of approval and non-approval; Converting the probability ratio into the scores of the risk factors and the total score of the target object; Conducting an effectiveness test on the risk sub-model and the approval result sub-model, and establishing the machine learning approval model when the test passes.

6. The method according to claim 5, wherein The extracting input feature variables according to the identity information of the target object includes: Extracting feature variables according to the identity information of the target object to form a feature wide table, and performing a preliminary screening on the feature wide table to form a candidate feature pool; Performing a secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables; Performing chi-square binning on the effective feature variables, adjusting the coordinate point threshold of the binning, and determining the input feature variables according to the evidence weight of each bin.

7. An information processing apparatus for a target object, characterized in that, The device includes: A cooperation request module, configured to receive a cooperation request initiated by a target object, and obtain the identity information of the target object based on the cooperation request; An expert scoring module, configured to approve the target object by using a pre-constructed expert scoring model, and obtain the scores of the risk factors and the total approval score of the target object; An expert approval module, configured to receive the first approval result of the target object based on the scores of the risk factors and the total approval score; the first approval result includes approval and non-approval; A sample module, configured to label the target object with a first approval result of non-approval, and store the identity information of the target object, the scores of the risk factors, the total approval score, the first approval result, and the label of the target object with a first approval result of non-approval in a sample database; A model approval module, configured to obtain sample data from the sample database, construct a machine learning approval model according to the sample data, and approve the target object by using the machine learning approval model. The machine learning approval model is used to output the second approval result of the target object and the scores of the risk factors of the target object; Among them, the model approval module includes: A sample extraction module, configured to obtain sample data of the first approval result of the target object from the sample database. The sample data includes the first approval result of the target object, the scores of the risk factors of the target object, and the label of the target object with a first approval result of non-approval; A model establishment module, configured to establish a risk sub-model based on the identity information of the target object and the scores of risk factors in the sample data, and establish an approval result sub-model based on the identity information of the target object and the first approval result in the sample data. The risk sub-model and the approval result sub-model constitute the machine learning approval model; the risk sub-model corresponds to the risk factor one by one.

8. The device according to claim 7, characterized in that, The device further includes: A forced screening module, configured to perform forced rule screening on the target object according to the identity information of the target object.

9. The device according to claim 7, characterized in that, The expert scoring module includes: An expert setting unit, configured to receive the risk factors set for the target object, the algorithm rules set for the risk factors, and the weights set for the risk factors. A scoring calculation unit, configured to establish the expert scoring model according to the risk factors, the algorithm rules set for the risk factors, and the weights set for the risk factors. The expert scoring model is used to output the scores of the risk factors of the target object and the total approval score according to the identity information of the target object.

10. The device according to claim 9, characterized in that, The sample module further includes: A labeling unit, configured to label the target object with unapproved approval, select the label of the reason for the unapproved approval of the target object on the labeling list, sort the labels according to a preset rule, and store the labels and label order in the sample database. An optimization unit, configured to optimize and update the expert scoring model according to the labels and label order of the target object.

11. The device according to claim 7, characterized in that, The model establishment module includes: A sample processing unit, configured to extract the input model feature variables according to the identity information of the target object. A regression algorithm unit, configured to calculate the input model feature variables based on a linear regression algorithm to obtain the probability ratio of approval and disapproval. A scoring output unit, configured to convert the probability ratio into the scores of the risk factors of the target object and the total score. An inspection unit, configured to perform validity inspection on the risk sub-model and the approval result sub-model. In the case of passing the inspection, establish the machine learning approval model.

12. The device according to claim 11, characterized in that, The sample processing unit includes: A preliminary screening sub-unit, configured to extract feature variables according to the identity information of the target object to form a feature wide table, and perform preliminary screening on the feature wide table to form a candidate feature pool. A secondary screening sub-unit, configured to perform secondary screening on the feature variables in the candidate feature pool to obtain effective feature variables. An input model feature variable unit, configured to perform chi-square binning on the effective feature variables, adjust the coordinate point threshold of the binning, and determine the input model feature variables according to the weight of evidence of each bin.

13. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images