CN113052604A - Object detection method, device, equipment and storage medium - Google Patents
Object detection method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113052604A CN113052604A CN202110288269.4A CN202110288269A CN113052604A CN 113052604 A CN113052604 A CN 113052604A CN 202110288269 A CN202110288269 A CN 202110288269A CN 113052604 A CN113052604 A CN 113052604A
- Authority
- CN
- China
- Prior art keywords
- node
- nodes
- risk
- detected
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Abstract
The application provides an object detection method, device, equipment and storage medium, and belongs to the technical field of risk safety. The method comprises the following steps: generating a bipartite graph corresponding to an object to be detected, wherein the bipartite graph comprises a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing the object to be detected, each second node is used for representing an operating device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes executes operation on the operating device represented by the second nodes; determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph; and determining a risk object in the object to be detected according to the risk value of each first node. The method and the device can improve the accuracy and timeliness of risk judgment of the object to be detected.
Description
Technical Field
The present application relates to the field of risk security technologies, and in particular, to a method, an apparatus, a device, and a storage medium for object detection.
Background
In some fields, risk detection is often required, for example, in the field of transactions, there is a need to detect fraud and other fraudulent activities in some transaction accounts.
Currently, when an account is detected, whether the account has fraud or other illegal behaviors is judged mainly based on transaction information of the account in a certain time and times.
However, the current method may result in lower accuracy and timeliness of the detection result.
Disclosure of Invention
The application aims to provide an object detection method, an object detection device, object detection equipment and a storage medium, which can improve the accuracy and timeliness of risk judgment of an object to be detected.
The embodiment of the application is realized as follows:
in one aspect of the embodiments of the present application, an object detection method is provided, including:
generating a bipartite graph corresponding to an object to be detected, wherein the bipartite graph comprises a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing the object to be detected, each second node is used for representing an operating device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes executes operation on the operating device represented by the second nodes;
determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph;
and determining a risk object in the object to be detected according to the risk value of each first node.
Optionally, determining a risk object in the object to be detected according to the risk value of each first node includes:
and predicting the first node with the risk value larger than the preset value by adopting a binary model to obtain a risk object in the object to be detected.
Optionally, before the first node with the risk value larger than the preset value is predicted by using a binary classification model to obtain a risk object in the object to be detected, the method includes:
obtaining a plurality of historical samples, wherein each historical sample comprises a risk record of a first node;
and (4) obtaining the characteristics in the historical sample by adopting a sliding window method and pre-training to obtain a two-classification model.
Optionally, determining a risk value of each first node by using a label propagation algorithm based on the bipartite graph, including:
determining a risk initial value of each first node in the bipartite graph;
and determining the risk value of each first node by adopting a label propagation algorithm by taking the risk initial value of each first node, the number of second nodes connected with the first nodes and the weight corresponding to the second nodes connected with the first nodes as parameters, wherein the weight corresponding to the second nodes connected with the first nodes is obtained based on the operation data of the first nodes on the second nodes connected with the first nodes.
Optionally, before determining the risk value of each first node by using a label propagation algorithm with the initial risk value of each first node, the number of second nodes having a connection with the first node, and a weight corresponding to the second nodes having a connection with the first node as parameters, the method further includes:
and analyzing the node degree of the second node, and deleting the second node with the node degree larger than the preset node degree in the bipartite graph and the corresponding connecting line.
Optionally, after determining a risk object in the object to be detected according to the risk value of each first node, the method further includes:
updating the risk initial value of each first node in the bipartite graph based on the risk value of the risk object to obtain a new risk initial value;
and determining a new risk value of each first node by adopting a label propagation algorithm based on the new risk initial value.
Optionally, the method further comprises:
and training a binary model by adopting the new risk value of each first node to obtain a new binary model.
In another aspect of the embodiments of the present application, an object detecting apparatus is provided, including: the system comprises a generating module, a risk determining module and an object determining module;
the generation module is used for generating a bipartite graph corresponding to the object to be detected, the bipartite graph comprises a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing the object to be detected, each second node is used for representing an operation device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes executes operation on the operation device represented by the second nodes;
the risk determination module is used for determining the risk value of each first node by adopting a label propagation algorithm based on the bipartite graph;
and the object determining module is used for determining a risk object in the object to be detected according to the risk value of each first node.
Optionally, the object determining module is specifically configured to perform prediction processing on the first node with the risk value larger than the preset value by using a binary model, so as to obtain a risk object in the object to be detected.
Optionally, the apparatus further comprises: a model building module; the model establishing module is used for acquiring a plurality of historical samples, and each historical sample comprises a risk record of a first node; and (4) obtaining the characteristics in the historical sample by adopting a sliding window method and pre-training to obtain a two-classification model.
Optionally, the risk determining module is specifically configured to determine a risk initial value of each first node in the bipartite graph;
and determining the risk value of each first node by adopting a label propagation algorithm by taking the risk initial value of each first node, the number of second nodes connected with the first nodes and the weight corresponding to the second nodes connected with the first nodes as parameters, wherein the weight corresponding to the second nodes connected with the first nodes is obtained based on the operation data of the first nodes on the second nodes connected with the first nodes.
Optionally, the risk determining module is further configured to perform node degree analysis on the second node, and delete the second node and the corresponding connection line in the bipartite graph, where the node degree is greater than the preset node degree.
Optionally, the risk determining module is further configured to update a risk initial value of each first node in the bipartite graph based on the risk value of the risk object to obtain a new risk initial value; and determining a new risk value of each first node by adopting a label propagation algorithm based on the new risk initial value.
Optionally, the model establishing module is further configured to train a two-class model by using the new risk value of each first node, so as to obtain a new two-class model.
In another aspect of the embodiments of the present application, there is provided a computer device, including: the object detection method comprises a memory and a processor, wherein a computer program capable of running on the processor is stored in the memory, and when the computer program is executed by the processor, the steps of the object detection method are realized.
In another aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the above-mentioned object detection method.
The beneficial effects of the embodiment of the application include:
in an object detection method, an object detection device, an object detection apparatus, and a storage medium provided in the embodiments of the present application, a bipartite graph corresponding to an object to be detected may be generated, where the bipartite graph includes a plurality of first nodes and a plurality of second nodes, the first nodes are connected to the second nodes by connecting lines, each first node is used to represent an object to be detected, each second node is used to represent an operation device, and the connecting lines between the first nodes and the second nodes are used to represent that the object to be detected represented by the first nodes has performed an operation on the operation device represented by the second nodes; determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph; the risk object in the object to be detected can be determined according to the risk value of each first node. The risk value is determined by adopting a label propagation algorithm, the transaction information of the object to be detected within a certain time and times does not need to be referred, and further the risk value of the object to be detected can be determined in time under the condition that the transaction information does not exist; or when it is determined that the operating device corresponding to a certain node has a risk, the object to be detected corresponding to the node connected to the node corresponding to the operating device is also used as a risk object, so that the accuracy of judging whether the object to be detected is the risk object can be improved. That is to say, the timeliness and the accuracy of judging whether the object to be detected is a risk object can be improved through the label propagation algorithm.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a schematic structural diagram of a bipartite graph according to an embodiment of the present application;
fig. 2 is a first schematic flowchart of an object detection method according to an embodiment of the present application;
fig. 3 is a schematic flowchart illustrating a second object detection method according to an embodiment of the present application;
fig. 4 is a third schematic flowchart of an object detection method according to an embodiment of the present application;
fig. 5 is a fourth schematic flowchart of an object detection method according to an embodiment of the present application;
fig. 6 is a fifth flowchart of an object detection method according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of an object detection apparatus according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
In the description of the present application, it is noted that the terms "first", "second", "third", and the like are used merely for distinguishing between descriptions and are not intended to indicate or imply relative importance.
The following specifically explains a specific structure of the bipartite graph and a connection relationship of each element in the bipartite graph provided in the embodiments of the present application.
Fig. 1 is a schematic structural diagram of a bipartite graph according to an embodiment of the present application, and please refer to fig. 1, where the bipartite graph includes a plurality of first nodes and a plurality of second nodes, the first nodes are connected to the second nodes through connecting lines, each first node is used to represent an object to be detected, each second node is used to represent an operating device, and the connecting lines between the first nodes and the second nodes are used to represent that the object to be detected represented by the first nodes has performed an operation on the operating device represented by the second nodes.
Optionally, the bipartite graph may include a plurality of first nodes and a plurality of second nodes, each first node may be connected to one or more second nodes through a connection line, and correspondingly, each second node may also be connected to one or more first nodes through a connection line. When the application is applied to different scenes, the meaning of the first node and the second node can be different. For example, when the node is applied to the field of credit card transactions, the object to be detected may refer to a credit card account, and correspondingly, the object to be detected represented by the first node may be the credit card account, and specifically, the credit card account may be identified by information such as an identification card identifier, a credit card identifier, a mobile phone number, and the like, which is not limited herein; accordingly, the operating device characterized by the second node may be an identification of the electronic device for collection, for example: any information such as computer identification, identification of POS (point of sale), identification of merchant, etc., and is not limited herein.
For example: in fig. 1, 4 first nodes and 4 second nodes are taken as an example, and a first node on the left side in fig. 1 is connected to a first second node and a second node on the right side through a connection line, so that it can be determined that an object to be detected corresponding to the first node can perform an operation on an operation device corresponding to the first second node and an operation device corresponding to the second node.
The following specifically explains a specific implementation process of the object detection method provided in the embodiment of the present application.
Fig. 2 is a first schematic flowchart of an object detection method according to an embodiment of the present application, please refer to fig. 2, where the method includes:
s210: and generating a bipartite graph corresponding to the object to be detected.
Optionally, an execution main body of the method according to the embodiment of the present application may be a computer device, and the computer device may generate a bipartite graph corresponding to an object to be detected by accessing, downloading, transmitting, acquiring, or actively drawing, where the object to be detected may be a plurality of objects to be detected that need to perform object detection, and each object to be detected is identified as one of the first nodes in the bipartite graph.
S220: and determining the risk value of each first node by adopting a label propagation algorithm based on the bipartite graph.
Alternatively, the label propagation algorithm may be a risk value calculation method based on a bipartite graph, and specifically may be to settle a risk value of each first node according to each first node in the bipartite graph and a plurality of second nodes having a connection relationship with the first node. The risk value may be a probability that the node may have a transaction risk in a transaction process, specifically, may be represented in a percentage manner, or may be represented by a specific numerical value, which is not limited herein, and may be set correspondingly according to a judgment requirement of a user.
Optionally, according to a connection relationship between each first node and each second node in the bipartite graph, a risk value of each first node may be calculated and determined.
S230: and determining a risk object in the object to be detected according to the risk value of each first node.
Optionally, after the risk value of each first node is determined, a preset model may be adopted to specifically determine whether each object to be detected in the plurality of first nodes is a risk object according to the magnitude of the risk value. And further all risk objects in the object to be detected can be determined.
The risk object can be an object with transaction risk in the object to be detected.
In the object detection method provided by the embodiment of the application, a bipartite graph corresponding to an object to be detected can be generated, the bipartite graph includes a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing an object to be detected, each second node is used for representing an operation device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes is operated on the operation device represented by the second nodes; determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph; the risk object in the object to be detected can be determined according to the risk value of each first node. The risk value is determined by adopting a label propagation algorithm, the transaction information of the object to be detected within a certain time and times does not need to be referred, and further the risk value of the object to be detected can be determined in time under the condition that the transaction information does not exist; or when it is determined that the operating device corresponding to a certain node has a risk, the object to be detected corresponding to the node connected to the node corresponding to the operating device is also used as a risk object, so that the accuracy of judging whether the object to be detected is the risk object can be improved. That is to say, the timeliness and the accuracy of judging whether the object to be detected is a risk object can be improved through the label propagation algorithm.
Optionally, determining a risk object in the object to be detected according to the risk value of each first node includes:
and predicting the first node with the risk value larger than the preset value by adopting a binary model to obtain a risk object in the object to be detected.
Alternatively, the classification model may be a classification model trained in advance, and specifically, the first nodes may be classified into two types, i.e., risky and risk-free types by the classification model, and the probability that each first node is classified into the two types may be calculated respectively. In addition, before the classification, the first nodes can be further screened, further determination is performed according to the risk value obtained in the step S220, the first nodes with the risk value higher than the preset value are screened, and prediction processing of the secondary classification model is performed.
Optionally, after the prediction processing is performed through the two classification models, the probabilities of the risk types of each first node in the prediction result may be ranked, and the first node with the probability higher than the preset probability is taken out as the risk node, so that the object to be detected corresponding to the risk node is the risk object.
According to the object detection method, the risk object in the object to be detected is obtained by adopting the two-classification model to predict the first node with the risk value larger than the preset value, after the node with the risk is preliminarily determined through the label propagation mode, the risk object is further more accurately determined from the preliminarily determined risk node through the two-classification model, the accuracy of determining the risk object can be further improved, and the probability of misjudgment is reduced.
Another specific implementation procedure of the object detection method provided in the embodiment of the present application is specifically explained below.
Fig. 3 is a second schematic flow chart of the object detection method provided in the embodiment of the present application, and please refer to fig. 3, where a binary model is used to perform prediction processing on a first node with a risk value greater than a preset value, before a risk object in an object to be detected is obtained, the method includes:
s310: a plurality of historical samples are obtained.
And each history sample comprises a risk record of the first node.
Optionally, the risk record may be a record used for recording risk transactions that have occurred historically in a certain first node, and specifically, these historical samples may be obtained from past transaction records, and each historical sample records a risk record of a different first node.
S320: and (4) obtaining the characteristics in the historical sample by adopting a sliding window method and pre-training to obtain a two-classification model.
Optionally, after the history sample is obtained, each feature in the history sample may be obtained by using a sliding window method, where the sliding window method may be to set a certain window length, obtain a certain part of features in the history sample under the window length for determination, and slide the sliding window to select another part of features in the history sample for determination.
Optionally, the derivative characteristics in each historical sample can be determined by a sliding window method, so that the coverage of the sample is wider, a preset model is trained by the plurality of historical samples, and a target model is obtained, wherein the target model is the binary classification model.
In the object detection method provided by the embodiment of the application, the two classification models are obtained by obtaining the plurality of historical samples and adopting a sliding window method to obtain the characteristics in the historical samples and pre-training the characteristics, so that the two classification models can be more stable, the risk type of the first node can be more accurately determined, and accordingly, the accuracy of determining the risk object can be improved.
Next, another specific implementation process of the object detection method provided in the embodiment of the present application will be explained.
Fig. 4 is a third schematic flowchart of an object detection method according to an embodiment of the present application, please refer to fig. 4, which determines a risk value of each first node by using a label propagation algorithm based on a bipartite graph, including:
s410: and determining the initial value of the risk of each first node in the bipartite graph.
Alternatively, the initial value of risk for each first node may be a preset size, for example: preset by the user.
S420: and determining the risk value of each first node by adopting a label propagation algorithm by taking the risk initial value of each first node, the number of second nodes connected with the first nodes and the weight corresponding to the second nodes connected with the first nodes as parameters.
And obtaining the weight corresponding to the second node with the connecting line with the first node based on the operation data of the first node on the second node with the connecting line with the first node.
Optionally, after determining the initial risk value of the first node, the risk value of the first node may be further calculated, where the specific calculation formula is as follows:
where score (P) is the risk value of the first node P, neighbor (P) represents all the corresponding second nodes of the first node P, and n is the number of second nodes corresponding to the first node P, where Σ weight is a normalization factor in order to make the sum of the probabilities 1. weight (i) represents a weight between the first node p and the second node i, and weight (i) is determined by the operation data of the second node i to the first node p, for example: data of the transfer amount, etc.; score (i) is the risk value of a second node adjacent to the first node.
Optionally, the risk value of each second node may be obtained by calculating according to the risk initial value of the first node, the specific calculation manner is similar to the formula, which is not described herein, after the risk value of each second node is determined, the score (i) is determined, and the risk value score (P) of the first node P may be obtained by calculating according to the calculation formula.
Optionally, in an actual application transaction, since the first node and the second node are very large in magnitude, a plurality of multivariate linear equations can be obtained by using the formula, and it is difficult to directly solve the equations by using a gaussian elimination method, and therefore, a plurality of equations after the formula is substituted by the equations can be specifically solved by using a matrix iteration calculation method.
Optionally, before determining the risk value of each first node by using a label propagation algorithm with the initial risk value of each first node, the number of second nodes having a connection with the first node, and a weight corresponding to the second nodes having a connection with the first node as parameters, the method further includes:
and analyzing the node degree of the second node, and deleting the second node with the node degree larger than the preset node degree in the bipartite graph and the corresponding connecting line.
Optionally, the node degree may be a value used for representing the connection number of one node and other nodes, and the greater the number of other nodes connected to one node, the higher the node degree of the node is, the node degrees of all the second nodes may be obtained, and the size relationship between the node degree of each second node and the preset node degree may be compared, and the size of the preset node degree may be the size of the node degree that is correspondingly set by the user according to actual needs, which is not limited specifically herein.
Optionally, a second node with a node degree greater than the preset node degree may be deleted, and accordingly, a plurality of wires corresponding to the second node may be deleted.
In the object detection method provided by the embodiment of the application, the second node with the node degree higher than the preset node degree in the bipartite graph and the corresponding connecting line are deleted by analyzing the node degree of the second node, so that the second node with the node degree higher than the node degree in the bipartite graph can be removed, and the accuracy of risk calculation is affected because the second node with the higher node degree has a poor effect when the second node is calculated, therefore, the accuracy of calculating the node risk value can be improved by removing the second node through the method, and the accuracy of determining the risk object can be further improved.
Next, a further specific implementation process of the object detection method provided in the embodiment of the present application is explained.
Fig. 5 is a fourth schematic flowchart of the object detection method according to the embodiment of the present application, please refer to fig. 5, where after the risk object in the object to be detected is determined according to the risk value of each first node, the method further includes:
s510: and updating the risk initial value of each first node in the bipartite graph based on the risk value of the risk object to obtain a new risk initial value.
Optionally, after the risk value of each risk object is determined, the risk object may be subjected to marking processing, where the marking processing specifically may be to identify the risk value of the first node corresponding to the risk object as the risk value of the risk object, and then return the risk value to the bipartite graph, so as to update the risk initial value of the first node in the bipartite graph, and update the risk initial value to a new risk initial value.
Optionally, the marking process may be marking by a computer device according to a preset algorithm or marking may be implemented by a service expert in a manner of inputting to a computer after considering various factors comprehensively, and a specific marking manner may be selected according to actual requirements without limitation.
S520: and determining a new risk value of each first node by adopting a label propagation algorithm based on the new risk initial value.
Optionally, after determining the new risk initial value, the label propagation algorithm may be used again to perform the calculation, specifically, the new risk initial value of each first node, the number of second nodes connected to the first node, and the weight corresponding to the second nodes connected to the first node may be used as parameters, and the label propagation algorithm is used to determine the risk value of each first node, where a specific calculation formula is the same as that described above, and details are not repeated here.
Optionally, the method further comprises:
and training a binary model by adopting the new risk value of each first node to obtain a new binary model.
Optionally, after determining the new risk value of each first node, the nodes may also be input into a binary model for training, and then a new binary model is obtained through training, and accordingly, the new binary model may be adopted to further determine the risk of the first node with risk calculated by the label propagation algorithm, where a specific determination process is similar to the foregoing process, and is not described herein again.
In the object detection method provided in the embodiment of the application, the risk initial value of each first node in the bipartite graph is updated based on the risk value of the risk object to obtain a new risk initial value, and then the new risk value of each first node is determined based on the new risk initial value by adopting a label propagation algorithm, and a binary model is trained by adopting the new risk value of each first node to obtain a new binary model. After the risk value is determined, the new risk initial value is obtained through updating, the new binary model is obtained through updating, the whole detection process can be closed, the accuracy of a label propagation algorithm and the accuracy of the binary model can be further improved, and the accuracy of determining whether the object to be detected is a risk object can be further improved.
The identification process of the object detection method and its implementation process in the present application are explained below by specific examples.
Fig. 6 is a fifth flowchart illustrating an object detection method according to an embodiment of the present application, please refer to fig. 6, where the method includes:
s610: and performing primary detection on the object to be detected by adopting a label propagation algorithm, and determining a suspected risk object in the object to be detected.
S620: and further determining the risk of the suspected risk object by adopting a binary classification model to obtain a high-risk object.
S630: and marking the high-risk object to obtain the risk object.
Alternatively, in practical applications, the objects to be detected are generally millions of units, the suspected risk objects are generally thousands of units, and the high-risk objects are generally hundreds of units, and after the marking process, the risk objects can be determined from the high-risk objects.
In the object detection method provided in the embodiment of the application, through the two stages of sequential risk determination, the risk object can be determined from millions of units of objects to be detected, accordingly, the application range of the method can be increased, and the higher accuracy of risk identification can be kept in a large number of scenes of the objects to be detected.
The following describes apparatuses, devices, and storage media for executing the object detection method provided by the present application, and specific implementation procedures and technical effects thereof are referred to above, and will not be described again below.
Fig. 7 is a schematic structural diagram of an object detection apparatus according to an embodiment of the present application, please refer to fig. 7, which includes: a generation module 710, a risk determination module 720, an object determination module 730;
a generating module 710, configured to generate a bipartite graph corresponding to an object to be detected, where the bipartite graph includes a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used to represent an object to be detected, each second node is used to represent an operating device, and the connecting lines between the first nodes and the second nodes are used to represent that the object to be detected represented by the first nodes has performed an operation on the operating device represented by the second nodes;
a risk determining module 720, configured to determine a risk value of each first node by using a label propagation algorithm based on the bipartite graph;
and the object determining module 730 is configured to determine a risk object in the object to be detected according to the risk value of each first node.
Optionally, the object determining module 730 is specifically configured to perform prediction processing on the first node with the risk value larger than the preset value by using a binary model, so as to obtain a risk object in the object to be detected.
Optionally, the apparatus further comprises: a model building module 740; a model building module 740, configured to obtain a plurality of historical samples, where each historical sample includes a risk record of a first node; and (4) obtaining the characteristics in the historical sample by adopting a sliding window method and pre-training to obtain a two-classification model.
Optionally, the risk determining module 720 is specifically configured to determine a risk initial value of each first node in the bipartite graph; and determining the risk value of each first node by adopting a label propagation algorithm by taking the risk initial value of each first node, the number of second nodes connected with the first nodes and the weight corresponding to the second nodes connected with the first nodes as parameters, wherein the weight corresponding to the second nodes connected with the first nodes is obtained based on the operation data of the first nodes on the second nodes connected with the first nodes.
Optionally, the risk determining module 720 is further configured to perform node degree analysis on the second node, and delete the second node and the corresponding connection line in the bipartite graph, where the node degree is greater than the preset node degree.
Optionally, the risk determining module 720 is further configured to update a risk initial value of each first node in the bipartite graph based on the risk value of the risk object, so as to obtain a new risk initial value; and determining a new risk value of each first node by adopting a label propagation algorithm based on the new risk initial value.
Optionally, the model building module 740 is further configured to train a binary model by using the new risk value of each first node, so as to obtain a new binary model.
The above-mentioned apparatus is used for executing the method provided by the foregoing embodiment, and the implementation principle and technical effect are similar, which are not described herein again.
These above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or one or more microprocessors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), among others. For another example, when one of the above modules is implemented in the form of a Processing element scheduler code, the Processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).
Fig. 8 is a schematic structural diagram of a computer device according to an embodiment of the present application, and referring to fig. 8, the computer device includes: the object detection method comprises a memory 810 and a processor 820, wherein a computer program capable of running on the processor 820 is stored in the memory 810, and when the computer program is executed by the processor 820, the steps of the object detection method are realized.
In another aspect of the embodiments of the present application, there is also provided a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the above-mentioned object detection method.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, a division of a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute some steps of the methods according to the embodiments of the present invention. And the aforementioned storage medium includes: a U disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. An object detection method, comprising:
generating a bipartite graph corresponding to an object to be detected, wherein the bipartite graph comprises a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing the object to be detected, each second node is used for representing an operating device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes executes operation on the operating device represented by the second nodes;
determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph;
and determining a risk object in the object to be detected according to the risk value of each first node.
2. The method according to claim 1, wherein the determining a risk object in the objects to be detected according to the risk value of each first node comprises:
and predicting the first node with the risk value larger than the preset value by adopting a binary model to obtain a risk object in the object to be detected.
3. The method according to claim 2, wherein before the first node with the risk value larger than the preset value is subjected to prediction processing by using a binary model to obtain the risk object in the object to be detected, the method comprises:
obtaining a plurality of historical samples, wherein each historical sample comprises a risk record of a first node;
and obtaining the characteristics in the historical samples by adopting a sliding window method and pre-training to obtain the two classification models.
4. The method of claim 1, wherein determining a risk value for each of the first nodes using a label propagation algorithm based on the bipartite graph comprises:
determining a risk initial value of each first node in the bipartite graph;
and determining the risk value of each first node by adopting the label propagation algorithm by taking the risk initial value of each first node, the number of second nodes connected with the first nodes and the weight corresponding to the second nodes connected with the first nodes as parameters, wherein the weight corresponding to the second nodes connected with the first nodes is obtained based on the operation data of the first nodes on the second nodes connected with the first nodes.
5. The method of claim 4, wherein before determining the risk value of each first node using the label propagation algorithm with the initial risk value of each first node, the number of second nodes connected to the first node, and the weight corresponding to the second nodes connected to the first node as parameters, the method further comprises:
and analyzing the node degree of the second node, and deleting the second node with the node degree larger than the preset node degree in the bipartite graph and the corresponding connecting line.
6. The method according to claim 5, wherein after determining the risk object in the objects to be detected according to the risk value of each of the first nodes, the method further comprises:
updating the risk initial value of each first node in the bipartite graph based on the risk value of the risk object to obtain a new risk initial value;
and determining a new risk value of each first node by adopting the label propagation algorithm based on the new risk initial value.
7. The method of claim 6, wherein the method further comprises:
and training a binary model by adopting the new risk value of each first node to obtain a new binary model.
8. An object detecting apparatus, characterized by comprising: the system comprises a generating module, a risk determining module and an object determining module;
the generation module is used for generating a bipartite graph corresponding to the object to be detected, the bipartite graph comprises a plurality of first nodes and a plurality of second nodes, the first nodes are connected with the second nodes through connecting lines, each first node is used for representing the object to be detected, each second node is used for representing an operating device, and the connecting lines between the first nodes and the second nodes are used for representing that the object to be detected represented by the first nodes executes operation on the operating device represented by the second nodes;
the risk determination module is used for determining a risk value of each first node by adopting a label propagation algorithm based on the bipartite graph;
and the object determining module is used for determining a risk object in the object to be detected according to the risk value of each first node.
9. A computer device, comprising: memory in which a computer program is stored which is executable on the processor, and a processor which, when executing the computer program, carries out the steps of the method according to any one of the preceding claims 1 to 7.
10. A computer-readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110288269.4A CN113052604A (en) | 2021-03-17 | 2021-03-17 | Object detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110288269.4A CN113052604A (en) | 2021-03-17 | 2021-03-17 | Object detection method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113052604A true CN113052604A (en) | 2021-06-29 |
Family
ID=76513242
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110288269.4A Pending CN113052604A (en) | 2021-03-17 | 2021-03-17 | Object detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113052604A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117061252A (en) * | 2023-10-12 | 2023-11-14 | 杭州智顺科技有限公司 | Data security detection method, device, equipment and storage medium |
-
2021
- 2021-03-17 CN CN202110288269.4A patent/CN113052604A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117061252A (en) * | 2023-10-12 | 2023-11-14 | 杭州智顺科技有限公司 | Data security detection method, device, equipment and storage medium |
| CN117061252B (en) * | 2023-10-12 | 2024-03-12 | 杭州智顺科技有限公司 | Data security detection method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009174B (en) | Risk recognition model training method and device and server | |
| Cao et al. | Titant: Online real-time transaction fraud detection in ant financial | |
| CN109360089B (en) | Loan risk prediction method and device | |
| CN102667775B (en) | Method for training and using a classification model with association rule models | |
| CN111796957B (en) | Transaction abnormal root cause analysis method and system based on application log | |
| CN112837069A (en) | Block chain and big data based secure payment method and cloud platform system | |
| CN108596759A (en) | loan application information detecting method and server | |
| CN114549001A (en) | Method and device for training risk transaction recognition model and recognizing risk transaction | |
| CN112529575A (en) | Risk early warning method, equipment, storage medium and device | |
| CN112561320A (en) | Training method of mechanism risk prediction model, mechanism risk prediction method and device | |
| CN110310123B (en) | Risk judging method and device | |
| CN112990989B (en) | Value prediction model input data generation method, device, equipment and medium | |
| CN111179051A (en) | Financial target customer determination method and device and electronic equipment | |
| CN112581271B (en) | Merchant transaction risk monitoring method, device, equipment and storage medium | |
| CN114091684A (en) | Method and device for enhancing interpretability of service result | |
| CN114202336A (en) | Risk behavior monitoring method and system in financial scene | |
| CN111245815B (en) | Data processing method and device, storage medium and electronic equipment | |
| CN113052604A (en) | Object detection method, device, equipment and storage medium | |
| CN111951008A (en) | Risk prediction method and device, electronic equipment and readable storage medium | |
| CN111582647A (en) | User data processing method and device and electronic equipment | |
| CN115392787A (en) | Enterprise risk assessment method, device, equipment, storage medium and program product | |
| CN110570301B (en) | Risk identification method, device, equipment and medium | |
| CN113435900A (en) | Transaction risk determination method and device and server | |
| CN114722954A (en) | Content exception handling method and device for evaluation information | |
| CN112116358A (en) | Transaction fraud prediction method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |