CN114612887A

CN114612887A - Bill abnormity detection method, device, equipment and computer readable storage medium

Info

Publication number: CN114612887A
Application number: CN202111021032.6A
Authority: CN
Inventors: 徐超; 李晓雯; 赵瑞辉; 多慧娟; 郑建光; 倪剑文; 郑冶枫
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-09-01
Filing date: 2021-09-01
Publication date: 2022-06-10
Anticipated expiration: 2041-09-01
Also published as: CN114612887B

Abstract

The application discloses a bill abnormity detection method, device and equipment and a computer readable storage medium, and belongs to the technical field of data processing. The method comprises the following steps: acquiring a document to be processed of a target user, wherein the document to be processed is a document which is not subjected to resource restoration; acquiring a historical document of a target user, wherein the historical document is a document subjected to resource restoration; and determining an abnormal detection result corresponding to the document to be processed based on the document to be processed and the historical document. According to the method, an abnormity detection rule does not need to be established manually, so that the abnormity detection process of the document to be processed is more reasonable, and the abnormity detection result of the document to be processed is higher in accuracy.

Description

Bill abnormity detection method, device, equipment and computer readable storage medium

Technical Field

The embodiment of the application relates to the technical field of data processing, in particular to a bill abnormity detection method, device and equipment and a computer readable storage medium.

Background

With the continuous development of data processing technology, the types of documents are increasing, and some documents are documents capable of resource restoration (reimbursement), such as medical insurance documents. In order to make the resource restoration of the document more accurate, before the resource restoration of the document, the document needs to be detected abnormally, and then the document passing the abnormal detection needs to be restored.

In the related art, taking a document to be processed as an example of a medical insurance document, the process of performing anomaly detection on the document to be processed is as follows: and determining a resource reduction value of the user corresponding to the document to be processed and a diagnosis and treatment department corresponding to the document to be processed. And responding to the fact that the resource restoration value is smaller than the resource restoration threshold value and the diagnosis and treatment department is in the scope of the diagnosis and treatment department, and determining that the bill to be processed is not an abnormal bill. Otherwise, determining the document to be processed as an abnormal document.

However, the resource reduction threshold and the diagnosis and treatment department range involved in the process of carrying out the anomaly detection on the document to be processed are manually established based on experience, so that the anomaly detection process of the document to be processed is not reasonable enough, and the accuracy of the anomaly detection result of the document to be processed is low.

Disclosure of Invention

The embodiment of the application provides a bill abnormity detection method, a bill abnormity detection device, equipment and a computer readable storage medium, which can be used for solving the problems in the related technology. The technical scheme is as follows:

in one aspect, an embodiment of the present application provides a method for detecting a document abnormality, where the method includes:

acquiring a document to be processed of a target user, wherein the document to be processed is a document which is not subjected to resource restoration;

acquiring a historical document of the target user, wherein the historical document is a document subjected to resource restoration;

and determining an abnormal detection result corresponding to the to-be-processed bill based on the to-be-processed bill and the historical bill.

On the other hand, an embodiment of the present application provides a document abnormality detection apparatus, the apparatus includes:

the acquiring module is used for acquiring a document to be processed of a target user, wherein the document to be processed is a document which is not subjected to resource restoration;

the acquisition module is used for acquiring the historical document of the target user, wherein the historical document is a document subjected to resource restoration;

and the determining module is used for determining an abnormal detection result corresponding to the to-be-processed bill based on the to-be-processed bill and the historical bill.

In a possible implementation manner, the obtaining module is configured to obtain a first feature vector based on the to-be-processed document; acquiring a second feature vector based on the historical document;

the determining module is used for inputting the first feature vector and the second feature vector into an abnormal probability prediction model to obtain a reference probability;

and determining an abnormal detection result corresponding to the bill to be processed based on the reference probability.

In a possible implementation manner, the determining module is configured to determine, based on the reference probability, an abnormal probability corresponding to the to-be-processed document;

in response to the fact that the abnormal probability is larger than a probability threshold value, determining that an abnormal detection result of the to-be-processed bill is a first result, wherein the first result is used for indicating that the to-be-processed bill cannot be subjected to resource restoration;

and in response to the fact that the abnormal probability is smaller than the probability threshold, determining that the abnormal detection result of the to-be-processed bill is a second result, wherein the second result is used for indicating that the to-be-processed bill can be subjected to resource restoration.

In a possible implementation manner, the obtaining module is further configured to obtain a facial image of the target user in response to the anomaly probability being equal to the probability threshold;

acquiring first feature data corresponding to the face image;

acquiring second characteristic data corresponding to the target user;

the determining module is further configured to determine a matching degree between the first feature data and the second feature data;

responding to the fact that the matching degree is smaller than a matching threshold value, and determining that the abnormal detection result of the to-be-processed bill is the first result;

and determining that the abnormal detection result of the to-be-processed bill is the second result in response to the matching degree not being smaller than the matching threshold.

In a possible implementation manner, the obtaining module is configured to obtain first data corresponding to the to-be-processed document; calling a feature vector acquisition model to process the first data to obtain a first feature vector corresponding to the first data;

acquiring second data corresponding to the historical document; calling the feature vector acquisition model to process the second data to obtain an initial feature vector corresponding to the second data; acquiring an initial feature matrix based on the initial feature vector corresponding to the second data, wherein the initial feature matrix comprises at least one initial feature vector corresponding to the second data; calling a target loss function to process the initial characteristic matrix to obtain a target characteristic matrix; and taking the feature vector included by the target feature matrix as a second feature vector corresponding to second data included by the initial feature matrix.

In one possible implementation, the apparatus further includes:

the generating module is used for generating an information obtaining request based on the user identification of the target user, wherein the information obtaining request carries the user identification, and the information obtaining request is used for obtaining the historical document of the target user;

the sending module is used for sending the information acquisition request to first electronic equipment, and the first electronic equipment stores the history bills of each user;

and the receiving module is used for receiving the history bill of the target user returned by the first electronic equipment.

In one possible implementation, the apparatus further includes:

and the processing module is used for processing the document to be processed according to the abnormal detection result corresponding to the document to be processed.

On the other hand, an embodiment of the present application provides an electronic device, where the electronic device includes a processor and a memory, where the memory stores at least one program code, and the at least one program code is loaded and executed by the processor, so that the electronic device implements any one of the above-mentioned bill abnormality detection methods.

In another aspect, a computer-readable storage medium is provided, in which at least one program code is stored, and the at least one program code is loaded and executed by a processor, so as to enable a computer to implement any one of the above-mentioned bill abnormality detection methods.

In another aspect, a computer program or a computer program product is provided, in which at least one computer instruction is stored, and the at least one computer instruction is loaded and executed by a processor, so as to enable a computer to implement any one of the above-mentioned document abnormality detection methods.

The technical scheme provided by the embodiment of the application at least has the following beneficial effects:

according to the technical scheme, before the document to be processed is subjected to resource restoration, the document to be processed is subjected to abnormity detection based on the document to be processed and the historical document of which the user has performed resource restoration, and an abnormity detection result of the document to be processed is determined. When the method is used for carrying out abnormity detection on the document to be processed, an abnormity detection rule does not need to be established manually, so that the abnormity detection process of the document to be processed is more reasonable, and the accuracy of an abnormity detection result of the document to be processed can be improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a schematic diagram of an implementation environment of a document anomaly detection method according to an embodiment of the present application;

FIG. 2 is a flowchart of a document anomaly detection method according to an embodiment of the present application;

FIG. 3 is a diagram illustrating a knowledge-graph corresponding to a target user according to an embodiment of the present disclosure;

FIG. 4 is a flowchart illustrating an embodiment of the present application for determining an abnormal detection result of a document to be processed;

FIG. 5 is an architecture diagram of a document anomaly detection method according to an embodiment of the present application;

FIG. 6 is a schematic structural diagram of a document anomaly detection device according to an embodiment of the present application;

fig. 7 is a schematic structural diagram of a terminal device according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of a server according to an embodiment of the present application.

Detailed Description

To make the objects, technical solutions and advantages of the present application more clear, the following detailed description of the embodiments of the present application will be made with reference to the accompanying drawings.

Fig. 1 is a schematic diagram of an implementation environment of a document abnormality detection method provided in an embodiment of the present application, and as shown in fig. 1, the implementation environment includes: an electronic device 101. The bill abnormity detection method provided by the embodiment of the application can be executed by the electronic equipment 101. Illustratively, the electronic device 101 may include any one of a terminal device or a server.

The terminal device may be at least one of a smart phone, a game console, a desktop computer, a tablet computer, an e-book reader, an MP3(Moving Picture Experts Group Audio Layer III, motion Picture Experts compression standard Audio Layer 3) player, an MP4(Moving Picture Experts Group Audio Layer IV, motion Picture Experts compression standard Audio Layer 4) player, and a laptop computer.

The terminal device may refer to one of a plurality of terminal devices, and this embodiment is only illustrated by the terminal device. Those skilled in the art will appreciate that the number of terminal devices described above may be greater or fewer. For example, the number of the terminal devices may be only one, or the number of the terminal devices may be tens or hundreds, or may be more, and the number of the terminal devices and the device types are not limited in the embodiment of the present application.

The server may be one server, or a server cluster formed by multiple servers, or any one of a cloud computing platform and a virtualization center, which is not limited in this embodiment of the present application. The server can be in communication connection with the terminal device through a wired network or a wireless network. The server may have functions of data processing, data storage, data transceiving, and the like. Of course, the server may also have other functions, which are not limited in this embodiment of the application.

Based on the foregoing implementation environment, an embodiment of the present application provides a method for detecting a document abnormality, which may be executed by the electronic device 101 in fig. 1, taking a flowchart of the method for detecting a document abnormality provided in the embodiment of the present application shown in fig. 2 as an example. As shown in fig. 2, the method comprises the steps of:

in step 201, a to-be-processed document of a target user is obtained, where the to-be-processed document is a document that is not subjected to resource restoration.

In the exemplary embodiment of the application, an image pickup device is installed and operated in the electronic device, and the image pickup device is used for scanning the graphic code of the target user to obtain the graphic code of the target user. After the electronic equipment acquires the graphic code of the target user, the graphic code is analyzed to obtain the user identification of the target user. The electronic equipment stores the to-be-processed document and the corresponding relation between the to-be-processed document and the user identification. The electronic equipment obtains the to-be-processed document of the target user based on the user identification of the target user and the corresponding relation between the to-be-processed document and the user identification. The number of the documents to be processed of each user is only one.

The resource recovery is used for deducting part of resources in the resource numerical value corresponding to the document, namely when the user transfers the resources to the document, part of the resources in the resource numerical value corresponding to the document are not required to be transferred by the user, and the user only needs to transfer the resources except for part of the resources in the resource numerical value corresponding to the document. The bill which is not subjected to resource restoration is used for indicating the bill which is not subjected to deduction of part of resources in the resource value corresponding to the bill.

For example, the resource value corresponding to the document is 100 yuan, and the resource that can be deducted from the resource value corresponding to the document is 60 yuan, so that when the document is subjected to resource transfer, the resource that the user needs to transfer is 40 yuan.

The graphic code may be a Bar code (Bar code) or a two-dimensional code (2-dimensional Bar code), and the embodiment of the present application does not limit the style of the graphic code. The graphic code of the target user is related to the type of the document to be processed, and when the type of the document to be processed is the medical insurance document, the graphic code of the target user is the medical insurance graphic code. The camera device may be a camera of an electronic device, an infrared scanning device, or another device, which is not limited in this embodiment of the present application.

The process of acquiring the graphic code of the target user by the electronic equipment comprises the following steps: the client for displaying the graphic code is installed and operated in the terminal device of the target user, and the client may be any type of client, which is not limited in the embodiment of the present application. And when the terminal equipment of the target user receives the display instruction, the terminal equipment of the target user displays the graphic code of the target user. The electronic equipment scans the graphic code of the target user by calling the camera device of the electronic equipment to obtain the graphic code of the target user.

The location of the electronic equipment is also related to the type of the document to be processed, and the location of the electronic equipment is a location supporting resource restoration. For example, when the type of the document to be processed is a medical insurance document, the location where the electronic device is located is a medical institution, and the medical institution may be a hospital, a pharmacy, or other locations supporting resource restoration, which is not limited in the embodiment of the present application.

The user identifier of the target user is an identifier capable of uniquely representing the target user, and the user identifier of the target user is not limited in the embodiment of the application. When the type of the document to be processed is a medical insurance document, the user identifier of the target user can be a medical insurance electronic certificate of the target user. When the user identifier of the target user is the medical insurance electronic certificate of the target user, the medical insurance electronic certificate may be composed of numbers, letters, or a combination of numbers and letters, which is not limited in the embodiment of the present application. One medical insurance electronic certificate can correspond to only one user.

In a possible implementation manner, the process of analyzing the graphic code by the electronic device to obtain the user identifier of the target user is as follows: and analyzing the graphic code by adopting a graphic code analyzer to obtain an identifier corresponding to the graphic code, and taking the identifier corresponding to the graphic code as the user identifier of the target user. The graphic code parser is any device capable of parsing a graphic code, and the embodiment of the present application does not limit this.

For example, the graphic code analyzer is used for analyzing the graphic code to obtain the medical insurance electronic certificate corresponding to the graphic code, and the medical insurance electronic certificate is used as the user identifier of the target user.

In one possible implementation, the graphic code displayed on the client of the terminal device of the target user is updated periodically. For example, the graphics code is updated every three minutes, the client at time 13: 00, starting to display the graphic code, and if the graphic code fails at the moment 13:03, refreshing a new graphic code on the terminal equipment of the user, wherein the new graphic code is still used for acquiring the user identifier of the target user.

It should be noted that, the above-mentioned update cycle of the graphic code is only three minutes, and is not used to limit the update cycle of the graphic code. The update period of the graphic code may be longer or shorter, which is not limited in the embodiment of the present application.

In step 202, a history document of the target user is obtained, and the history document is a document subjected to resource restoration.

In a possible implementation manner, after the user identifier of the target user is obtained in step 201, an information obtaining request is generated based on the user identifier of the target user. The information obtaining request carries a user identifier of a target user, the information obtaining request is used for obtaining a historical document of the target user, the historical document is a document subjected to resource restoration, namely the historical document is a document subjected to deduction of partial resources in a resource value corresponding to the document when the document is subjected to resource transfer, and the historical document is a document subjected to reimbursement. The electronic equipment is in communication connection with first electronic equipment through a wired network or a wireless network, and the first electronic equipment stores history bills of each user. For example, the first electronic device is an electronic device (e.g., an electronic device of a medical insurance institution) dedicated to resource recovery of the document. The electronic equipment sends the information acquisition request to the first electronic equipment. The first electronic device stores the historical documents of each user and the corresponding relation between the user identification and the historical documents of the user. The first electronic equipment receives an information acquisition request sent by the electronic equipment, analyzes the information acquisition request and obtains a user identifier of a target user carried in the information acquisition request. And acquiring the historical document of the target user based on the user identification of the target user and the corresponding relation between the user identification and the historical document of the user. The first electronic device sends the history document of the target user to the electronic device.

In one possible implementation, the history document of the target user may be one or more. When the history documents of the target user are multiple, the first electronic device may sequentially send each history document to the electronic device, and the electronic device sequentially receives the history documents sent by the first electronic device, that is, the electronic device obtains the history documents of the target user. The first electronic device may also package a plurality of history documents to obtain a compressed packet, and send the compressed packet to the electronic device. The electronic device receives the compressed packet and decompresses the compressed packet to obtain the historical document of the target user, namely the electronic device obtains the historical document of the target user.

In step 203, an anomaly detection result corresponding to the document to be processed is determined based on the document to be processed and the history document.

In a possible implementation manner, based on the document to be processed and the history document, a process of determining an abnormal detection result corresponding to the document to be processed is as follows: acquiring a first feature vector based on a document to be processed; acquiring a second feature vector based on the historical document; inputting the first feature vector and the second feature vector into an abnormal probability prediction model to obtain a reference probability; and determining an abnormal detection result corresponding to the document to be processed based on the reference probability. The abnormal probability prediction model is used for determining the probability that the bill to be processed is an abnormal bill, and the abnormal bill is used for indicating the bill which cannot be subjected to resource restoration.

Based on the document to be processed, the process of obtaining the first feature vector comprises the following steps: acquiring first data corresponding to a document to be processed. And calling a feature vector acquisition model to process the first data to obtain a first feature vector corresponding to the first data. That is, the first data is input into the feature vector acquisition model, and the first feature vector corresponding to the first data is obtained based on the output result of the feature vector acquisition model.

The first data includes, but is not limited to, a position of an object corresponding to the to-be-processed document, a user corresponding to the to-be-processed document, a resource value corresponding to the to-be-processed document, and generation time of the to-be-processed document. If the document to be processed is a medical insurance document, the object corresponding to the document to be processed is a medical institution corresponding to the document to be processed, and the position of the object corresponding to the document to be processed is the position of the medical institution corresponding to the document to be processed. The user corresponding to the document to be processed is a user who has a doctor in the medical institution, the resource value corresponding to the document to be processed is a resource value of the document to be processed, which needs to be subjected to resource transfer, and the generation time of the document to be processed is the time for generating the document to be processed. Of course, the first data may also be other information, which is not limited in this embodiment of the application.

The feature vector acquisition model is used for acquiring a feature vector corresponding to the data. The feature vector obtaining model may be a Long Short-Term Memory network (LSTM) model, or may be another model capable of obtaining a feature vector, which is not limited in this embodiment of the present application.

Based on the historical documents, the process of obtaining the second feature vector comprises the following steps: and acquiring second data corresponding to the historical documents. And calling the feature vector acquisition model to process the second data to obtain an initial feature vector corresponding to the second data. And acquiring an initial characteristic matrix based on the initial characteristic vector corresponding to the second data, wherein the initial characteristic matrix comprises at least one initial characteristic vector corresponding to the second data. Calling a target loss function to process the initial characteristic matrix to obtain a target characteristic matrix; and taking the feature vector included by the target feature matrix as a second feature vector corresponding to second data included by the initial feature matrix.

The second data includes, but is not limited to, a location of an object corresponding to the history document, a behavior corresponding to the history document, a user corresponding to the history document, a sub-object (a visiting department in a medical institution) included in the object corresponding to the history document, a control user (a visiting doctor in the visiting department) of the sub-object corresponding to the history document, generation time of the history document, a resource value corresponding to the history document, and a resource restoration value corresponding to the history document. The actions corresponding to the history document include, but are not limited to, actions of purchasing medicine, outpatient service, hospitalization, and the like. The resource restoration value corresponding to the history document is a resource value which does not need the user to perform resource transfer when the history document is subjected to resource transfer (that is, a resource value which can be reimbursed when the history document is subjected to resource transfer).

Because the historical documents are the documents acquired from the first electronic device, and the documents stored in the first electronic device are the documents subjected to resource restoration, the number of the second data corresponding to the historical documents is more than the number of the first data corresponding to the documents to be processed.

The process of calling the feature vector acquisition model to process the second data to obtain the initial feature vector corresponding to the second data is consistent with the process of calling the feature vector acquisition model to process the first data to obtain the first feature vector corresponding to the first data, and details are not repeated here.

Illustratively, the first data corresponding to the pending documents of the target user are user a, medical institution a and 120 meta. The number of the history documents of the target user is two, the second data corresponding to the first history document is a user A, an outpatient service and a medical institution A, and the second data corresponding to the second history document is a user A, a hospitalization and a medical institution B.

Based on the initial feature vector corresponding to the second data, the process of obtaining the initial feature matrix is as follows: constructing a knowledge graph corresponding to the target user based on the second data, wherein the knowledge graph comprises at least one piece of knowledge, and each piece of knowledge comprises at least one piece of second data; and acquiring an initial characteristic matrix based on the initial characteristic vector and the knowledge graph corresponding to the second data. That is, the initial feature vector corresponding to the second data included in each piece of knowledge in the knowledge graph is formed into the initial feature matrix corresponding to the piece of knowledge.

Take the example that the initial feature matrix corresponding to knowledge includes three initial feature vectors. The initial feature matrix corresponding to knowledge is of the form (h, l, t). And h, l and t are initial feature vectors corresponding to second data included in the knowledge graph respectively, and h, l and t are three different initial feature vectors respectively.

Taking the second data as user a, an outpatient service, a medical institution a, a medicine purchasing and a medical institution B as examples, fig. 3 is a schematic diagram of a target user knowledge graph provided in the embodiment of the present application. In this fig. 3, two pieces of knowledge are shown, wherein one piece of knowledge is: user a-clinic-medical institution a, knowledge two is: user a-purchase-medical institution B.

Taking the initial feature vector corresponding to the user a as a vector M, the initial feature vector corresponding to the medical institution a as a vector N, the initial feature vector corresponding to the medical institution B as a vector O, the initial feature vector corresponding to the outpatient service as a vector P, and the initial feature vector corresponding to the medicine purchase as a vector Q as an example. The initial feature matrix corresponding to the first knowledge is (M, P, N), and the initial feature matrix corresponding to the second knowledge is (M, Q, O).

In a possible implementation manner, the process of calling a target loss function to process the initial feature matrix to obtain a target feature matrix is as follows: based on the initial feature matrix, a reference feature matrix is determined, the reference feature matrix being a feature matrix that is inconsistent with the initial feature matrix. And calling a target loss function to process the reference characteristic matrix and the initial characteristic matrix to obtain a target characteristic matrix.

For example, the second initial feature vector in the reference feature matrix is identical to the second initial feature vector in the initial feature matrix, the first initial feature vector in the reference feature matrix is not identical to the first initial feature vector in the initial feature matrix, and/or the third initial feature vector in the reference feature matrix is not identical to the third initial feature vector in the initial feature matrix.

Calling a target loss function to process the reference characteristic matrix and the initial characteristic matrix, wherein the process of obtaining the target characteristic matrix comprises the following steps: and calling a target loss function to process the reference characteristic matrix and the initial characteristic matrix to obtain a first loss value. And in response to the first loss value meeting the loss requirement, taking the initial feature matrix as a target feature matrix. And in response to the first loss value not meeting the loss requirement, adjusting the initial feature matrix to obtain a first feature matrix. And calling a target loss function to process the first characteristic matrix and the reference characteristic matrix to obtain a second loss value. And in response to the second loss value meeting the loss requirement, taking the first feature matrix as a target feature matrix. And responding to the fact that the second loss value still does not meet the loss requirement, and continuing to perform iterative processing until the feature matrix with the loss value meeting the loss requirement is obtained. And taking the feature matrix with the loss value meeting the loss requirement as a target feature matrix.

Wherein, the loss value meeting the loss requirement means that the loss value is less than the loss threshold. The loss threshold may be any value, and may be adjusted based on an application scenario, which is not limited in this embodiment of the present application.

And after the target characteristic matrix is determined, the characteristic vector included by the target characteristic matrix is used as a second characteristic vector corresponding to second data included by the initial characteristic matrix.

Illustratively, knowledge one is user a-clinic-medical institution a and knowledge one includes second data that is user a, clinic and medical institution a. The initial feature matrix corresponding to the knowledge-knowledge matrix is composed of the initial feature vector corresponding to the user a, the initial feature vector corresponding to the clinic, and the initial feature vector corresponding to the medical institution a, that is, the initial feature matrix is (M, P, N). The reference feature matrix is (M, P, O). And calling a target loss function to process the initial characteristic matrix and the reference characteristic matrix to obtain a first loss value. And in response to the first loss value meeting the loss requirement, taking the initial feature matrix as a target feature matrix, namely, taking the target feature matrix as (M, P, N). Therefore, M is the second feature vector corresponding to the user a, P is the second feature vector corresponding to the clinic, and N is the second feature vector corresponding to the medical institution a.

And in response to the first loss value not meeting the loss requirement, adjusting the initial feature matrix to obtain a first feature matrix (X, Y, Z), and calling a target loss function to process the first feature matrix and the reference feature matrix to obtain a second loss value. And in response to the second loss value meeting the loss requirement, taking the first feature matrix as a target feature matrix, namely, taking the target feature matrix as (X, Y, Z). Therefore, X is the second feature vector corresponding to the user a, Y is the second feature vector corresponding to the clinic, and Z is the second feature vector corresponding to the medical institution a.

In one possible implementation, the reference feature matrix is determined according to the following equation (1) based on the initial feature matrix.

S′_(h，l，t)＝{(h′，l，t)|h′∈E}∪{(h，l，t′)|t′∈F} (1)

In the above formula (1), S'_(h，l，t)For the set of reference feature matrices, h is the first initial feature vector in the initial feature matrix, l is the second initial feature vector in the initial feature matrix, and t is the third initial feature vector in the initial feature matrix. E is a set of initial feature vectors of all initial feature vectors included in the knowledge-graph except for a first initial feature vector of the initial feature matrix, and F is a set of initial feature vectors of all initial feature vectors included in the knowledge-graph except for a third initial feature vector of the initial feature matrix.

Take the target loss function as maxmargin (maximum profit) loss function as an example. The following formula (2) is a calculation process of the loss value provided in the embodiment of the present application.

In the above formula (2), L is a loss value, (h, L, t) is an initial feature matrix, S is a set of initial feature matrices corresponding to knowledge included in the knowledge map, (h ', L, t ') is a reference feature matrix, S '_(h，l，t)For a set of reference feature matrices, γ is a margin hyper-parameter greater than 0, d (h + l, t) is the distance of the sum of the initial feature vector h and the initial feature vector 1 from the initial feature vector t, and d (h '+ l, t') is the distance of the sum of the initial feature vector h 'and the initial feature vector l from the initial feature vector t'.

It should be noted that the target loss function may also be other loss functions, and the above only takes the target loss function as max margin loss function as an example, and is not used to limit the type of the target loss function.

In a possible implementation manner, after the second feature vector corresponding to the second data is acquired, the second feature vector corresponding to the second data and the user identifier of the target user are correspondingly stored in the electronic device. When a new document to be processed exists in the target user, when the new document to be processed is subjected to abnormity detection, the second feature vector corresponding to the second data is directly extracted, the second feature vector is not required to be obtained based on the historical document of the target user again, and the obtaining time of the second feature vector can be saved.

In a possible implementation manner, after a first feature vector and a second feature vector are obtained, the first feature vector and the second feature vector are input into an abnormal probability prediction model to obtain an abnormal detection result corresponding to a document to be processed, namely the first feature vector and the second feature vector are input into the abnormal probability prediction model to obtain a reference probability; and determining an abnormal detection result corresponding to the document to be processed based on the reference probability.

In one possible implementation, the anomaly probability prediction model includes at least one of a first probability prediction model, a second probability prediction model, a third probability prediction model, and a fourth probability prediction model. In response to the anomaly probability prediction model comprising the first probability prediction model, the second probability prediction model, the third probability prediction model, and the fourth probability prediction model, the reference probability comprises the first probability, the second probability, the third probability, and the fourth probability. The first probability is obtained by inputting the first feature vector and the second feature vector into a first probability prediction model. The second probability is obtained by inputting the first feature vector and the second feature vector into a second probability prediction model. The third probability is obtained by inputting the first feature vector and the second feature vector into a third probability prediction model. The fourth probability is obtained by inputting the first feature vector and the second feature vector into a fourth probability prediction model.

Illustratively, the anomaly probability prediction model is at least One of an eliptic Envelope model, a One Class SVM (support vector machine-like) model, an Isolation Forest model, and a Robust Covariance model. Of course, the anomaly probability prediction model may be other models, which is not limited in the embodiment of the present application.

The One Class SVM model maps the second feature vector to a high-dimensional space, and a linear classifier is learned in the high-dimensional space, so that most of the second feature vectors and the origin are respectively positioned at two sides of the classifier. When the first feature vector is processed, the probability that the first feature vector and the second feature vector are on different sides is judged, and the probability is used as a reference probability.

The Isolation Forest model is a non-parametric unsupervised model that treats samples that are sparsely distributed and far from a high density (feature vectors of documents) as anomalous samples (anomalous documents). By using a forest model to iteratively generate a classification surface, a feature space (a space containing all samples) is divided into a plurality of areas until each area only contains one sample (a feature vector of a receipt), abnormal samples are resolved by calculating the depth of each sample on each tree (area), and the samples with the depth lower than a certain threshold value are taken as the abnormal samples.

In a possible implementation manner, based on the reference probability, the process of determining the abnormal detection result corresponding to the document to be processed is as follows: and determining the abnormal probability corresponding to the bill to be processed based on the reference probability. And determining an abnormal detection result corresponding to the document to be processed based on the abnormal probability.

Based on the reference probability, the process of determining the abnormal probability corresponding to the document to be processed comprises the following steps: acquiring a weight parameter corresponding to the abnormal probability prediction model, and determining the abnormal probability corresponding to the document to be processed based on the weight parameter corresponding to the abnormal probability prediction model and the reference probability.

In a possible implementation manner, based on the weight parameter and the reference probability corresponding to the abnormal probability prediction model, the abnormal probability P corresponding to the document to be processed is determined according to the following formula (3):

P＝A*α+B*β+C*γ+D*δ(3)

in the formula (3), P is an abnormal probability corresponding to the to-be-processed document, a is a first probability, B is a second probability, C is a third probability, D is a fourth probability, α is a weight parameter corresponding to the first probabilistic prediction model, β is a weight parameter corresponding to the second probabilistic prediction model, γ is a weight parameter corresponding to the third probabilistic prediction model, and δ is a weight parameter corresponding to the fourth probabilistic prediction model.

Illustratively, the first probability is 0.7, the second probability is 0.3, the third probability is 0.6, and the fourth probability is 0.8. The weight parameter corresponding to the first probabilistic predictive model is 0.25, the weight parameter corresponding to the second probabilistic predictive model is 0.25, the weight parameter corresponding to the third probabilistic predictive model is 0.25, and the weight parameter corresponding to the fourth probabilistic predictive model is 0.25. Based on the above formula (2), the anomaly probability of the document to be processed is obtained as P0.7 × 0.25+0.3 × 0.25+0.6 × 0.25+0.8 × 0.25 — 0.6.

It should be noted that the weighting parameter corresponding to each probabilistic prediction model may be any value, and the sum of the weighting parameters corresponding to each probabilistic prediction model may be 1 or may not be 1, which is not limited in the embodiment of the present application.

In a possible implementation manner, based on the anomaly probability, the process of determining the anomaly detection result corresponding to the document to be processed is as follows: and responding to the abnormal probability being larger than the probability threshold value, and determining that the abnormal detection result of the to-be-processed document is a first result, wherein the first result is used for indicating that the to-be-processed document cannot be subjected to resource restoration, namely the to-be-processed document is the document which cannot be reimbursed. And responding to the fact that the abnormal probability is smaller than the probability threshold value, determining that the abnormal detection result of the to-be-processed bill is a second result, wherein the second result is used for indicating that the to-be-processed bill can be subjected to resource restoration, namely the to-be-processed bill is a bill capable of being reimbursed.

Illustratively, the probability threshold is 0.5, and the corresponding abnormal probability of the to-be-processed document is 0.6. The abnormal probability corresponding to the document to be processed is greater than the probability threshold, so that the abnormal detection result of the document to be processed is determined to be the first result, that is, the document to be processed cannot be subjected to resource restoration.

For another example, the probability threshold is 0.7, and the anomaly probability corresponding to the document to be processed is 0.6. And determining that the abnormal detection result of the document to be processed is a second result, namely the document to be processed can be subjected to resource restoration.

It should be noted that the probability threshold may be any value, and the probability threshold may be adjusted based on an application scenario, which is not limited in this embodiment of the application.

In response to the abnormal probability being equal to the probability threshold, face verification is required, and the process of face verification is as follows: acquiring a facial image of a target user; acquiring first feature data corresponding to a face image; and acquiring second characteristic data corresponding to the target user. And determining the matching degree between the first characteristic data and the second characteristic data, and determining that the abnormal detection result of the document to be processed is a first result in response to the matching degree being smaller than the matching threshold value, namely determining that the document to be processed cannot be subjected to resource restoration when the face verification result is failed. And responding to the matching degree not less than the matching threshold value, determining that the abnormal detection result of the document to be processed is a second result, namely determining that the document to be processed can be subjected to resource restoration when the face verification result is passed.

When the abnormal probability is equal to the probability threshold, the electronic device displays prompt information, the prompt information is used for indicating to acquire a facial image of the target user, and the content of the prompt information is any content. The prompt message may be a text content displayed on the electronic device or an audio content, which is not limited in the embodiment of the present application. The camera device of the electronic equipment can collect images, the camera device of the electronic equipment collects the face images of the target users in response, and the electronic equipment extracts the features of the face images of the target users to obtain first feature data corresponding to the face images.

The feature extraction method may be a geometric feature-based method, a feature face method, a local feature method, a specific face subspace method, a hidden markov model method, or the like, and the feature extraction method is not limited in the embodiment of the present application. The feature data includes, but is not limited to, position information of facial feature points, a distance between eyes, a position between two corners of a mouth, a width of a forehead, and the like.

Illustratively, after the electronic device collects a facial image of the target user, the facial image of the target user is analyzed to obtain position information of the facial feature points of the target user, and the position information of the facial feature points of the target user is used as the first feature data.

The electronic device stores second characteristic data corresponding to each user and the corresponding relation between the user identification and the second characteristic data, and acquires the second characteristic data of the target user based on the user identification of the target user and the corresponding relation between the user identification and the second characteristic data.

The process of determining the matching degree between the first feature data and the second feature data is as follows: and determining a difference value between the first characteristic data and the second characteristic data, and taking the matching degree corresponding to the difference value as the matching degree between the first characteristic data and the second characteristic data. The matching degree between the first feature data and the second feature data may also be determined in other ways, which are not limited in the embodiments of the present application.

Illustratively, the probability threshold is 0.6, and the abnormal probability corresponding to the document to be processed is 0.6. And determining the matching degree of the first characteristic data and the second characteristic data of the target user because the probability threshold is consistent with the abnormal probability corresponding to the bill to be processed, and obtaining the matching degree of 80%. The matching threshold is 60%, and the matching degree is higher than the matching threshold, so that the abnormal detection result of the to-be-processed document is determined to be a second result, namely the to-be-processed document is the document capable of resource restoration.

It should be noted that the matching threshold may be any value, and the matching threshold may be adjusted based on an application scenario, and the value of the matching threshold is not limited in the embodiment of the present application.

Fig. 4 is a flowchart illustrating an abnormality detection result for determining a document to be processed according to an embodiment of the present application. In fig. 4, four abnormal probability prediction models are used, which are: a first probabilistic predictive model, a second probabilistic predictive model, a third probabilistic predictive model, and a fourth probabilistic predictive model. And inputting the first feature vector and the second feature vector into a first probability prediction model to obtain a first probability. And inputting the first feature vector and the second feature vector into a second probability prediction model to obtain a second probability. And inputting the first feature vector and the second feature vector into a third probability prediction model to obtain a third probability. And inputting the first feature vector and the second feature vector into a fourth probability prediction model to obtain a fourth probability. And determining the abnormal probability based on the first probability, the second probability, the third probability and the fourth probability. And determining an abnormal detection result of the document to be processed based on the abnormal probability.

In a possible implementation manner, after the abnormal detection result corresponding to the document to be processed is determined, the document to be processed is processed according to the abnormal detection result corresponding to the document to be processed.

In a possible implementation manner, based on an abnormality detection result corresponding to a document to be processed, there are two processing manners to process the document to be processed.

And in the first processing mode, the abnormal detection result of the document to be processed is the first result, and the document to be processed is not subjected to resource recovery.

In a possible implementation manner, in response to that the abnormal detection result of the to-be-processed document is the first result, the to-be-processed document is a document which cannot be subjected to resource restoration, that is, the to-be-processed document is an abnormal document and cannot be reimbursed, and then a first notification message is displayed. The first notification message is used for indicating the to-be-processed bill as a bill which cannot be subjected to resource restoration. The content of the first notification message may be any content, which is not limited in this embodiment of the present application.

And in the second processing mode, the abnormal detection result of the to-be-processed document is a second result, and the to-be-processed document is subjected to resource restoration.

In a possible implementation manner, in response to that the abnormal detection result of the to-be-processed document is the second result, that is, the to-be-processed document is the normal document, reimbursement can be performed to obtain a first resource value, where the first resource value is a restored resource value (reimbursed resource value) corresponding to the target user. And acquiring a second resource value based on the user identifier of the target user, wherein the second resource value is a resource reduction limit (reimbursement limit) corresponding to the target user. And responding to the situation that the first resource value is smaller than the second resource value, and acquiring a resource reduction ratio (reimbursement ratio) corresponding to the target user. And determining a third resource value based on the resource reduction proportion corresponding to the target user and the resource value corresponding to the document to be processed, wherein the third resource value is a reducible resource value in the resource values corresponding to the document to be processed. And performing resource restoration on the document to be processed according to the third resource value.

The process of obtaining the first resource value is as follows: and determining the historical documents with the occurrence time in the target time period in the historical documents of the target user. And acquiring a reduction resource numerical value corresponding to the historical document of which the occurrence time is in the target time period. And taking the sum of the restored resource values corresponding to the history bill with the occurrence time in the target time period as a first resource value.

The electronic device stores the resource reduction limit corresponding to the user and the corresponding relation between the user identification and the resource reduction limit corresponding to the user. And acquiring the resource reduction limit corresponding to the target user, namely acquiring a second resource value, based on the user identification of the target user and the corresponding relationship between the user identification and the resource reduction limit corresponding to the user.

It should be noted that the resource reduction quota is a resource reduction quota in the target time period. For example, if the first resource value is the resource value restored by the target user in 2021, the resource restoration quota is the resource restoration quota in 2021 year.

The target time period may be any time period, which is not limited in the embodiment of the present application.

As shown in the following table one, a table of a corresponding relationship between a user identifier and a resource reduction amount corresponding to a user is provided in an embodiment of the present application.

Watch 1

User identification	Resource recovery limit
		130421xxxxxxxx0001	10000
130421xxxxxxxx0002	5000
		130421xxxxxxxx0003	8000
130421xxxxxxxx0004	12000

As shown in the table I, the resource reduction quota corresponding to the user identified as 130421xxxxxxxx0001 is 10000 yuan. The resource reduction rate corresponding to the user with the user identification of 130421xxxxxxxx0002 is 5000 yuan. The resource reduction quota corresponding to the user with the user identification of 130421xxxxxxxx0003 is 8000 yuan. The resource reduction rate corresponding to the user with the user identification 130421xxxxxxxx0004 is 12000 yuan.

Illustratively, the user id of the target user is 130421xxxxxxxx0001, and the resource reduction quota corresponding to the target user is 10000 yuan. That is, the target subscriber has a reimburseable credit of 10000 dollars.

In a possible implementation manner, when the resource reduction ratio corresponding to the target user is obtained, the resource reduction ratio corresponding to the resource value corresponding to the document to be processed may be used as the resource reduction ratio of the target user. The resource reduction ratio of the target user may also be obtained in other manners, which is not limited in the embodiment of the present application.

Exemplarily, the resource value is 1300-30000 yuan, and the corresponding resource reduction ratio is 80%; the resource value is 30000-40000 yuan, and the corresponding resource reduction proportion is 90%; the resource value is 40000-100000 yuan, and the corresponding resource reduction ratio is 95%; the resource value exceeds 100000 yuan, and the corresponding resource reduction proportion is 80%. And if the resource value corresponding to the document to be processed is 2000 yuan, the resource reduction proportion of the target user is 80%.

In a possible implementation manner, before the to-be-processed document is subjected to resource restoration according to the third resource value, a fourth resource value is further determined based on the first resource value and the second resource value, and the fourth resource value is a resource restoration limit (a current reimburseable limit) currently available to the target user. And judging whether the third resource value is larger than the fourth resource value. And responding to the situation that the third resource value is not larger than the fourth resource value, and performing resource restoration on the to-be-processed document according to the third resource value. And responding to the situation that the third resource value is larger than the fourth resource value, and performing resource restoration on the to-be-processed document according to the fourth resource value.

Illustratively, the first resource value is 8000 yuan, the second resource value is 10000 yuan, the first resource value is smaller than the second resource value, the resource value corresponding to the document to be processed is 2000 yuan, and the resource reduction proportion of the target user is 80%. And determining the third resource value to be 2000 x 80 percent to 1600 yuan based on the resource reduction proportion and the resource value corresponding to the bill to be processed. And determining that the fourth resource value is 10000-8000-2000 yuan based on the first resource value and the second resource value, and performing resource restoration on the to-be-processed document according to the third resource value because the third resource value is smaller than the fourth resource value, namely, reimbursing 1600 yuan.

For another example, the first resource value is 8000 yuan, the second resource value is 10000 yuan, the first resource value is smaller than the second resource value, the resource value corresponding to the document to be processed is 5000 yuan, and the resource reduction ratio of the target user is 80%. And determining that the third resource value is 5000 x 80 percent to 4000 yuan based on the resource reduction proportion and the resource value corresponding to the bill to be processed. And determining that the fourth resource value is 10000-8000 yuan based on the first resource value and the second resource value, and performing resource restoration on the to-be-processed document according to the fourth resource value because the third resource value is greater than the fourth resource value, namely reimbursing for 2000 yuan.

In a possible implementation manner, in response to that the first resource value is equal to the second resource value, a second notification message is displayed, where the second notification message is used to indicate that the abnormal detection result of the to-be-processed document is the second result, but the resource reduction limit corresponding to the target user is used up. That is, the document to be processed is a reimburseable document, but the reimburseable amount of the target user is used up. The content of the second notification message may be any content, which is not limited in this embodiment of the present application.

In a possible implementation manner, after the document to be processed is processed based on the anomaly detection result, a fifth resource value may also be determined, where the fifth resource value is a resource value that the target user needs to transfer, that is, a part of resources that the target user needs to perform resource transfer in the document to be processed, and the fifth resource value is displayed.

The process of determining the value of the fifth resource is as follows:

and in response to the fact that the abnormal detection result of the to-be-processed bill is the first result, taking the resource value corresponding to the to-be-processed bill as a fifth resource value. That is, in response to the to-be-processed document being a document that cannot be subjected to resource restoration, the resource value corresponding to the to-be-processed document is taken as the fifth resource value.

Exemplarily, the document to be processed is a document which cannot be subjected to resource restoration, the resource value corresponding to the document to be processed is 5000 yuan, and the fifth resource value is 5000 yuan.

And in response to that the abnormal detection result of the to-be-processed document is the second result and the third resource value is not greater than the fourth resource value, taking the difference value between the resource value corresponding to the to-be-processed document and the third resource value as a fifth resource value. That is, if the document to be processed is a document capable of resource restoration, and the reimbursed resource value of the document to be processed is the third resource value, the difference value between the resource value corresponding to the document to be processed and the third resource value is taken as the fifth resource value.

Illustratively, the to-be-processed document is a document capable of performing resource restoration, the resource value corresponding to the to-be-processed document is 2000, the third resource value of the to-be-processed document that has been subjected to resource restoration is 1600 yuan, and then the fifth resource value is 2000-1600 yuan, which is 400 yuan.

And in response to the fact that the abnormal detection result of the to-be-processed bill is the second result and the third resource value is larger than the fourth resource value, taking the difference value between the resource value corresponding to the to-be-processed bill and the fourth resource value as a fifth resource value. That is, if the to-be-processed document is a document capable of resource restoration, and the reimbursed resource value of the to-be-processed document is the fourth resource value, the difference value between the resource value corresponding to the to-be-processed document and the fourth resource value is taken as the fifth resource value.

Illustratively, the to-be-processed document is a document capable of performing resource restoration, the resource value corresponding to the to-be-processed document is 5000 yuan, and the fourth resource value of the to-be-processed document that has been subjected to resource restoration is 2000 yuan, then the fifth resource value is 5000-2000-.

In a possible implementation manner, in response to that the abnormal detection result of the to-be-processed document is the second result, and the fifth resource value is determined, the resource value in the first account of the target user is obtained. It is determined whether the value of the resource in the first account of the target user is less than a fifth resource value. And in response to the resource value in the first account of the target user not being less than the fifth resource value, directly deducting the fifth resource value from the first account of the target user. And if the document to be processed is a medical insurance document, the first account is the medical insurance account of the target user.

And in response to the resource value in the first account of the target user being less than the fifth resource value, deducting all the resource values in the first account of the target user. And determining a sixth resource value based on the resource value in the first account of the target user and the fifth resource value, and deducting the sixth resource value from the second account of the target user. The second account of the target user is a bank card account of the target user or an account balance of the target user.

According to the method, before the document to be processed is subjected to resource restoration, the document to be processed is subjected to abnormity detection based on the document to be processed and the historical document of which the user has performed resource restoration, and an abnormity detection result of the document to be processed is determined. When the method is used for carrying out abnormity detection on the document to be processed, an abnormity detection rule does not need to be established manually, so that the abnormity detection process of the document to be processed is more reasonable, and the accuracy of an abnormity detection result of the document to be processed can be improved.

Fig. 5 is an architecture diagram of a document abnormality detection method according to an embodiment of the present application. As shown in fig. 5, the architecture diagram includes two parts, an offline stage and an online stage.

Wherein, the processing procedure of the off-line stage is as follows: and acquiring the historical documents stored in the medical insurance mechanism, and storing the historical documents in a data storage bank. And processing the data of the historical documents to obtain second data corresponding to the historical documents, and acquiring second feature vectors corresponding to the second data. The second feature vector is stored in a data store. The anomaly probability detection model is stored in a data store. The user representation is stored in a data store.

The treatment process of the online stage is as follows: and determining the user identification of the target user based on the medical insurance graphic code. And acquiring the to-be-processed document of the target user based on the user identification of the target user. And processing the data of the document to be processed to obtain first data corresponding to the document to be processed. And acquiring a first feature vector corresponding to the first data. And acquiring a second feature vector corresponding to the target user. And calling an abnormal probability detection model to process the first feature vector and the second feature vector to obtain abnormal probability. After the first feature vector is determined, the first feature vector and the document to be processed can be correspondingly stored in a data storage.

And when the abnormal probability is smaller than the probability threshold value, performing resource restoration on the document to be processed. And when the abnormal probability is larger than the probability threshold value, displaying the first notification message. When the abnormal probability is equal to the probability threshold, performing face verification, where the process of face verification is consistent with the process of determining the matching degree of the first feature data and the second feature data in step 204, and is not described herein again. And responding to the passing of the face verification, and performing resource recovery on the document to be processed. In response to the face verification failing, a first notification message is displayed.

Fig. 6 is a schematic structural diagram of a document abnormality detection apparatus according to an embodiment of the present application, and as shown in fig. 6, the apparatus includes:

the acquiring module 601 is configured to acquire a to-be-processed document of a target user, where the to-be-processed document is a document that is not subjected to resource restoration;

the acquiring module 601 is configured to acquire a history document of a target user, where the history document is a document subjected to resource restoration;

the determining module 602 is configured to determine, based on the to-be-processed document and the historical document, an abnormal detection result corresponding to the to-be-processed document.

In a possible implementation manner, the obtaining module 601 is configured to obtain a first feature vector based on a document to be processed; acquiring a second feature vector based on the historical document;

a determining module 602, configured to input the first feature vector and the second feature vector into an abnormal probability prediction model to obtain a reference probability;

and determining an abnormal detection result corresponding to the document to be processed based on the reference probability.

In a possible implementation manner, the determining module 602 is configured to determine, based on the reference probability, an abnormal probability corresponding to the to-be-processed document;

in response to the fact that the abnormal probability is larger than the probability threshold, determining that the abnormal detection result of the to-be-processed bill is a first result, wherein the first result is used for indicating that the to-be-processed bill cannot be subjected to resource restoration;

and responding to the fact that the abnormal probability is smaller than the probability threshold value, determining that the abnormal detection result of the to-be-processed bill is a second result, wherein the second result is used for indicating that the to-be-processed bill can be subjected to resource restoration.

In a possible implementation manner, the obtaining module 601 is further configured to obtain a facial image of the target user in response to the anomaly probability being equal to the probability threshold; acquiring first feature data corresponding to a face image; acquiring second characteristic data corresponding to a target user;

a determining module 602, further configured to determine a matching degree between the first feature data and the second feature data; responding to the fact that the matching degree is smaller than the matching threshold value, and determining that the abnormal detection result of the to-be-processed bill is a first result; and determining that the abnormal detection result of the to-be-processed bill is a second result in response to the matching degree not being smaller than the matching threshold.

In a possible implementation manner, the obtaining module 601 is configured to obtain first data corresponding to a to-be-processed document; calling a feature vector acquisition model to process the first data to obtain a first feature vector corresponding to the first data;

acquiring second data corresponding to the historical documents; calling a feature vector acquisition model to process the second data to obtain an initial feature vector corresponding to the second data; acquiring an initial feature matrix based on the initial feature vector corresponding to the second data, wherein the initial feature matrix comprises at least one initial feature vector corresponding to the second data; calling a target loss function to process the initial characteristic matrix to obtain a target characteristic matrix; and taking the feature vector included by the target feature matrix as a second feature vector corresponding to second data included by the initial feature matrix.

In one possible implementation, the apparatus further includes:

the generation module is used for generating an information acquisition request based on the user identifier of the target user, wherein the information acquisition request carries the user identifier and is used for acquiring the historical document of the target user;

the sending module is used for sending the information acquisition request to first electronic equipment, and the first electronic equipment is electronic equipment in which historical documents of various users are stored;

and the receiving module is used for receiving the history document of the target user returned by the first electronic equipment.

In one possible implementation, the apparatus further includes:

Before the device performs resource restoration on the document to be processed, the device performs anomaly detection on the document to be processed based on the document to be processed and the historical document of which the user has performed resource restoration, and determines an anomaly detection result of the document to be processed. When the document to be processed is detected abnormally, an abnormal detection rule does not need to be made manually, so that the abnormal detection process of the document to be processed is reasonable, and the accuracy of the abnormal detection result of the document to be processed can be improved.

It should be understood that, when the apparatus provided in fig. 6 implements its functions, it is only illustrated by the division of the functional modules, and in practical applications, the above functions may be distributed by different functional modules according to needs, that is, the internal structure of the apparatus is divided into different functional modules to implement all or part of the functions described above. In addition, the apparatus and method embodiments provided by the above embodiments belong to the same concept, and specific implementation processes thereof are described in the method embodiments for details, which are not described herein again.

Fig. 7 shows a block diagram of a terminal device 700 according to an exemplary embodiment of the present application. The terminal device 700 may be a portable mobile terminal such as: a smart phone, a tablet computer, an MP3(Moving Picture Experts Group Audio Layer III, motion video Experts compression standard Audio Layer 3) player, an MP4(Moving Picture Experts Group Audio Layer IV, motion video Experts compression standard Audio Layer 4) player, a notebook computer or a desktop computer. The terminal device 700 may also be referred to by other names such as user equipment, portable terminal, laptop terminal, desktop terminal, etc.

In general, the terminal device 700 includes: a processor 701 and a memory 702.

Processor 701 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so on. The processor 701 may be implemented in at least one hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array). The processor 701 may also include a main processor and a coprocessor, where the main processor is a processor for Processing data in an awake state, and is also called a Central Processing Unit (CPU); a coprocessor is a low power processor for processing data in a standby state. In some embodiments, the processor 701 may be integrated with a GPU (Graphics Processing Unit) which is responsible for rendering and drawing the content required to be displayed by the display screen. In some embodiments, the processor 701 may further include an AI (Artificial Intelligence) processor for processing computing operations related to machine learning.

Memory 702 may include one or more computer-readable storage media, which may be non-transitory. Memory 702 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in the memory 702 is used to store at least one instruction for execution by the processor 701 to implement the document anomaly detection method provided by the method embodiments of the present application.

In some embodiments, the terminal device 700 may further include: a peripheral interface 703 and at least one peripheral. The processor 701, the memory 702, and the peripheral interface 703 may be connected by buses or signal lines. Various peripheral devices may be connected to the peripheral interface 703 via a bus, signal line, or circuit board. Specifically, the peripheral device includes: at least one of a radio frequency circuit 704, a display screen 705, a camera assembly 706, an audio circuit 707, a positioning component 708, and a power source 709.

The peripheral interface 703 may be used to connect at least one peripheral device related to I/O (Input/Output) to the processor 701 and the memory 702. In some embodiments, the processor 701, memory 702, and peripheral interface 703 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 701, the memory 702, and the peripheral interface 703 may be implemented on a separate chip or circuit board, which is not limited in this embodiment.

The Radio Frequency circuit 704 is used for receiving and transmitting RF (Radio Frequency) signals, also called electromagnetic signals. The radio frequency circuitry 704 communicates with communication networks and other communication devices via electromagnetic signals. The rf circuit 704 converts an electrical signal into an electromagnetic signal to transmit, or converts a received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 704 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a subscriber identity module card, and so forth. The radio frequency circuitry 704 may communicate with other terminals via at least one wireless communication protocol. The wireless communication protocols include, but are not limited to: the world wide web, metropolitan area networks, intranets, generations of mobile communication networks (2G, 3G, 4G, and 5G), Wireless local area networks, and/or WiFi (Wireless Fidelity) networks. In some embodiments, the radio frequency circuit 704 may also include NFC (Near Field Communication) related circuits, which are not limited in this application.

The display screen 705 is used to display a UI (User Interface). The UI may include graphics, text, icons, video, and any combination thereof. When the display screen 705 is a touch display screen, the display screen 705 also has the ability to capture touch signals on or over the surface of the display screen 705. The touch signal may be input to the processor 701 as a control signal for processing. At this point, the display 705 may also be used to provide virtual buttons and/or a virtual keyboard, also referred to as soft buttons and/or a soft keyboard. In some embodiments, the display 705 may be one, disposed on the front panel of the terminal device 700; in other embodiments, the display 705 may be at least two, respectively disposed on different surfaces of the terminal device 700 or in a foldable design; in other embodiments, the display 705 may be a flexible display, disposed on a curved surface or on a folded surface of the terminal device 700. Even more, the display 705 may be arranged in a non-rectangular irregular pattern, i.e. a shaped screen. The Display 705 may be made of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), or the like.

The camera assembly 706 is used to capture images or video. Optionally, camera assembly 706 includes a front camera and a rear camera. Generally, a front camera is disposed at a front panel of the terminal, and a rear camera is disposed at a rear surface of the terminal. In some embodiments, the number of the rear cameras is at least two, and each rear camera is any one of a main camera, a depth-of-field camera, a wide-angle camera and a telephoto camera, so that the main camera and the depth-of-field camera are fused to realize a background blurring function, and the main camera and the wide-angle camera are fused to realize panoramic shooting and VR (Virtual Reality) shooting functions or other fusion shooting functions. In some embodiments, camera assembly 706 may also include a flash. The flash lamp can be a single-color temperature flash lamp or a double-color temperature flash lamp. The double-color-temperature flash lamp is a combination of a warm-light flash lamp and a cold-light flash lamp, and can be used for light compensation at different color temperatures.

The audio circuitry 707 may include a microphone and a speaker. The microphone is used for collecting sound waves of a user and the environment, converting the sound waves into electric signals, and inputting the electric signals to the processor 701 for processing or inputting the electric signals to the radio frequency circuit 704 to realize voice communication. For the purpose of stereo sound collection or noise reduction, a plurality of microphones may be provided at different positions of the terminal device 700. The microphone may also be an array microphone or an omni-directional pick-up microphone. The speaker is used to convert electrical signals from the processor 701 or the radio frequency circuit 704 into sound waves. The loudspeaker can be a traditional film loudspeaker or a piezoelectric ceramic loudspeaker. When the speaker is a piezoelectric ceramic speaker, the speaker can be used for purposes such as converting an electric signal into a sound wave audible to a human being, or converting an electric signal into a sound wave inaudible to a human being to measure a distance. In some embodiments, the audio circuitry 707 may also include a headphone jack.

The positioning component 708 is used to locate the current geographical position of the terminal device 700 to enable navigation or LBS (Location Based Service). The Positioning component 708 can be a Positioning component based on the Global Positioning System (GPS) in the united states, the beidou System in china, or the galileo System in russia.

The power supply 709 is used to supply power to various components in the terminal device 700. The power source 709 may be alternating current, direct current, disposable batteries, or rechargeable batteries. When the power source 709 includes a rechargeable battery, the rechargeable battery may be a wired rechargeable battery or a wireless rechargeable battery. The wired rechargeable battery is a battery charged through a wired line, and the wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery can also be used to support fast charge technology.

In some embodiments, the terminal device 700 further includes one or more sensors 170. The one or more sensors 170 include, but are not limited to: acceleration sensor 711, gyro sensor 712, pressure sensor 713, fingerprint sensor 714, optical sensor 715, and proximity sensor 716.

The acceleration sensor 711 may detect the magnitude of acceleration on three coordinate axes of the coordinate system established with the terminal device 700. For example, the acceleration sensor 711 may be used to detect components of the gravitational acceleration in three coordinate axes. The processor 701 may control the display screen 705 to display the user interface in a landscape view or a portrait view according to the gravitational acceleration signal collected by the acceleration sensor 711. The acceleration sensor 711 may also be used for acquisition of motion data of a game or a user.

The gyro sensor 712 may detect a body direction and a rotation angle of the terminal device 700, and the gyro sensor 712 may cooperate with the acceleration sensor 711 to acquire a 3D motion of the user with respect to the terminal device 700. From the data collected by the gyro sensor 712, the processor 701 may implement the following functions: motion sensing (such as changing the UI according to a user's tilting operation), image stabilization at the time of photographing, game control, and inertial navigation.

The pressure sensor 713 may be disposed on a side frame of the terminal device 700 and/or under the display 705. When the pressure sensor 713 is arranged on the side frame of the terminal device 700, the holding signal of the user to the terminal device 700 can be detected, and the processor 701 performs left-right hand recognition or shortcut operation according to the holding signal collected by the pressure sensor 713. When the pressure sensor 713 is disposed at a lower layer of the display screen 705, the processor 701 controls the operability control on the UI interface according to the pressure operation of the user on the display screen 705. The operability control comprises at least one of a button control, a scroll bar control, an icon control, and a menu control.

The fingerprint sensor 714 is used for collecting a fingerprint of a user, and the processor 701 identifies the identity of the user according to the fingerprint collected by the fingerprint sensor 714, or the fingerprint sensor 714 identifies the identity of the user according to the collected fingerprint. When the user identity is identified as a trusted identity, the processor 701 authorizes the user to perform relevant sensitive operations, including unlocking a screen, viewing encrypted information, downloading software, paying, changing settings, and the like. The fingerprint sensor 714 may be disposed on the front, back, or side of the terminal device 700. When a physical button or a vendor Logo is provided on the terminal device 700, the fingerprint sensor 714 may be integrated with the physical button or the vendor Logo.

The optical sensor 715 is used to collect the ambient light intensity. In one embodiment, the processor 701 may control the display brightness of the display screen 705 based on the ambient light intensity collected by the optical sensor 715. Specifically, when the ambient light intensity is high, the display brightness of the display screen 705 is increased; when the ambient light intensity is low, the display brightness of the display screen 705 is adjusted down. In another embodiment, processor 701 may also dynamically adjust the shooting parameters of camera assembly 706 based on the ambient light intensity collected by optical sensor 715.

A proximity sensor 716, also called a distance sensor, is typically provided on the front panel of the terminal device 700. The proximity sensor 716 is used to collect the distance between the user and the front surface of the terminal device 700. In one embodiment, when the proximity sensor 716 detects that the distance between the user and the front surface of the terminal device 700 gradually decreases, the processor 701 controls the display screen 705 to switch from the bright screen state to the dark screen state; when the proximity sensor 716 detects that the distance between the user and the front surface of the terminal device 700 gradually becomes larger, the processor 701 controls the display screen 705 to switch from the breath screen state to the bright screen state.

Those skilled in the art will appreciate that the configuration shown in fig. 7 does not constitute a limitation of terminal device 700 and may include more or fewer components than shown, or combine certain components, or employ a different arrangement of components.

Fig. 8 is a schematic structural diagram of a server according to an embodiment of the present application, where the server 800 may generate a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) 801 and one or more memories 802, where at least one program code is stored in the one or more memories 802, and is loaded and executed by the one or more processors 801 to implement the document anomaly detection method provided by the foregoing method embodiments. Of course, the server 800 may also have components such as a wired or wireless network interface, a keyboard, and an input/output interface, so as to perform input and output, and the server 800 may also include other components for implementing the functions of the device, which are not described herein again.

In an exemplary embodiment, there is also provided a computer readable storage medium having at least one program code stored therein, the at least one program code being loaded and executed by a processor to cause a computer to implement any of the above-mentioned document anomaly detection methods.

Alternatively, the computer-readable storage medium may be a Read-Only Memory (ROM), a Random Access Memory (RAM), a Compact Disc Read-Only Memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, and the like.

In an exemplary embodiment, there is also provided a computer program or computer program product having at least one computer instruction stored therein, the at least one computer instruction being loaded and executed by a processor to cause a computer to implement any of the above-described document anomaly detection methods.

It should be understood that reference to "a plurality" herein means two or more. "and/or" describes the association relationship of the associated object, indicating that there may be three relationships, for example, a and/or B, which may indicate: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship.

The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the advantages and disadvantages of the embodiments.

The above description is only exemplary of the application and should not be taken as limiting the application, and any modifications, equivalents, improvements and the like that are made within the spirit and principle of the application should be included in the protection scope of the application.

Claims

1. A document anomaly detection method is characterized by comprising the following steps:

2. The method according to claim 1, wherein the determining the abnormal detection result corresponding to the to-be-processed document based on the to-be-processed document and the history document comprises:

acquiring a first feature vector based on the bill to be processed;

acquiring a second feature vector based on the historical document;

inputting the first feature vector and the second feature vector into an abnormal probability prediction model to obtain a reference probability;

3. The method according to claim 2, wherein the determining the abnormal detection result corresponding to the to-be-processed document based on the reference probability comprises:

determining the abnormal probability corresponding to the bill to be processed based on the reference probability;

and in response to the abnormal probability being smaller than the probability threshold, determining that the abnormal detection result of the to-be-processed bill is a second result, wherein the second result is used for indicating that the to-be-processed bill can be subjected to resource restoration.

4. The method of claim 3, further comprising:

in response to the anomaly probability being equal to the probability threshold, obtaining a facial image of the target user;

acquiring first feature data corresponding to the face image;

acquiring second characteristic data corresponding to the target user;

determining a matching degree between the first characteristic data and the second characteristic data;

5. The method according to any of claims 2 to 4, wherein said obtaining a first feature vector based on said document to be processed comprises:

acquiring first data corresponding to the to-be-processed bill;

calling a feature vector acquisition model to process the first data to obtain a first feature vector corresponding to the first data;

the obtaining a second feature vector based on the historical document includes:

acquiring second data corresponding to the historical document;

calling the feature vector acquisition model to process the second data to obtain an initial feature vector corresponding to the second data;

acquiring an initial feature matrix based on the initial feature vector corresponding to the second data, wherein the initial feature matrix comprises at least one initial feature vector corresponding to the second data;

calling a target loss function to process the initial characteristic matrix to obtain a target characteristic matrix;

and taking the feature vector included by the target feature matrix as a second feature vector corresponding to second data included by the initial feature matrix.

6. The method according to any one of claims 1 to 4, wherein the obtaining of the history document of the target user comprises:

generating an information acquisition request based on the user identifier of the target user, wherein the information acquisition request carries the user identifier, and the information acquisition request is used for acquiring the historical document of the target user;

sending the information acquisition request to first electronic equipment, wherein the first electronic equipment stores history bills of each user;

and receiving the history document of the target user returned by the first electronic equipment.

7. The method according to any one of claims 1 to 4, wherein after determining the abnormal detection result corresponding to the to-be-processed document based on the to-be-processed document and the history document, the method further comprises:

and processing the bill to be processed according to the abnormal detection result corresponding to the bill to be processed.

8. A document anomaly detection apparatus, said apparatus comprising:

9. An electronic device, comprising a processor and a memory, wherein at least one program code is stored in the memory, and wherein the at least one program code is loaded into and executed by the processor, to cause the electronic device to implement a document anomaly detection method as claimed in any one of claims 1 to 7.

10. A computer readable storage medium having stored therein at least one program code, the at least one program code being loaded into and executed by a processor, to cause a computer to carry out a method of document anomaly detection according to any one of claims 1 to 7.

11. A computer program or computer program product having stored therein at least one computer instruction, the at least one computer instruction being loaded and executed by a processor, to cause a computer to implement a document anomaly detection method as claimed in any one of claims 1 to 7.