Zombie license intelligent identification method, device, equipment and storage medium
Through the K-fold cross-verification algorithm and random forest model, the identification of suspected zombie licenses is solved, and the problems of low recognition efficiency and low accuracy in the existing technology are improved, and the effect of food safety supervision is improved.
Patent Information
- Application Number
- CN202210255631.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-15
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2042-03-15
AI Technical Summary
In the prior art, zombie license identification is inefficient and has low accuracy, which makes it difficult for the government to effectively supervise and affect food safety.
The K-fold cross-validation algorithm and random forest model are used to build a data feature set by obtaining merchant information, marking normal and abnormal data, training a random forest model, filtering data features, and building a reconstruction model, which is used to identify suspected zombie licenses.
Improves the accuracy and efficiency of zombie license recognition, reduces data processing time, and enhances regulatory decision-making support capabilities.
Smart Images

Figure CN114638299B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, electronic device and computer-readable storage medium for intelligently identifying zombie licenses. Background Art
[0002] As provinces and cities continue to deepen the reform of the commercial registration system, they have introduced a number of measures such as "second approval" and "comprehensive window service", which have effectively improved the level of commercial registration and license processing services, and the number of commercial entities and licenses has increased rapidly. However, the number of "lost commercial entities" and "zombie licenses" that cannot be found, are out of supervision, and have no annual reports and tax records has also increased year by year. Not only have they occupied limited resources such as business names and trade names for a long time, but they have also brought hidden dangers to market transaction safety and increased market transaction costs. Especially in the field of food supervision, food operators have high mobility and high frequency of changes in business methods. In addition, the behavior of food operators' main qualifications being terminated according to law but the food business license has not been cancelled is particularly obvious, which poses a great hidden danger to food safety.
[0003] At present, "zombie licenses" can only be identified manually through field visits during the market supervision process, which makes it difficult for local governments to see the current situation and understand the bottom line in terms of resource allocation and overall decision-making, affecting the planning and arrangement of supervision tasks during the process, and bringing great challenges to supervision. The current zombie license identification method is inefficient and has the problem of low accuracy due to subjectivity or lack of experience in manual judgment. Summary of the invention
[0004] The present invention provides a method, a device and a computer-readable storage medium for intelligently identifying zombie licenses, the main purpose of which is to solve the problem of low efficiency and accuracy in identifying zombie licenses.
[0005] To achieve the above object, the present invention provides a method for intelligently identifying zombie licenses, comprising:
[0006] Acquire merchant information within a preset historical time period, and use the merchant information to construct a data feature set;
[0007] Based on the data feature set, a training set and a test set are constructed using a K-fold cross validation algorithm;
[0008] Marking normal data and abnormal data in the training set and the test set as positive and negative samples respectively;
[0009] Using the training set and the positive and negative samples to train a pre-constructed random forest model, obtaining a preliminarily trained random forest model and weights of each data feature, and screening the data features according to the weights to obtain retained data features;
[0010] Reconstructing the initially trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively training the reconstructed random forest model using the test set until a trained random forest model is obtained;
[0011] The trained random forest model is used to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
[0012] Optionally, constructing a data feature set using the merchant information includes:
[0013] Cleaning the merchant information, and performing multi-dimensional splitting on the cleaned merchant information to obtain multi-dimensional data;
[0014] Extracting information features of the plurality of dimensional data and association features between the dimensional data, and constructing a corresponding association relationship table using the information features and the association features;
[0015] The association relationship table is used to construct a data feature set.
[0016] Optionally, marking the normal data and the abnormal data in the training set and the test set as positive and negative samples, respectively, includes:
[0017] The relevant data in the training set and the test set that have no historical supervision evidence in the one-address-multiple-certificates are defined as abnormal data and marked as negative samples;
[0018] The relevant data of active merchants in the training set and the test set whose business addresses and business scopes are consistent with the licenses are defined as normal data and marked as positive samples.
[0019] Optionally, the pre-constructed random forest model is trained using the training set and the positive and negative samples to obtain a preliminarily trained random forest model and weights of each data feature, and the data features are screened according to the weights to obtain retained data features, including:
[0020] Using the decision layer in the pre-built random forest model, a root node is selected from the data features in the training set;
[0021] According to the root node, the classification layer in the pre-built random forest model is used to classify the remaining data features in the training set to obtain a plurality of leaf nodes and a random forest model that has been preliminarily trained;
[0022] According to the average accuracy of the positive and negative samples, the Gini coefficient of each leaf node is calculated using the activation function in the pre-built random forest model to obtain the weight of each data feature;
[0023] Determining whether the weight meets a preset condition;
[0024] If the weight does not meet the preset condition, adjust the parameters of the random forest model that has been preliminarily trained, delete the data features corresponding to the weights that do not meet the preset condition, and return to execute the step of classifying the remaining data features in the training set according to the root node using the classification layer in the pre-constructed random forest model;
[0025] If the weights meet the preset conditions, the retained data features are obtained.
[0026] Optionally, the using the decision layer in the pre-built random forest model to select a root node from the data features in the training set includes:
[0027] Calculating the information gain value of the data feature of the training data in the merchant information using the decision algorithm in the decision layer;
[0028] The data feature with the highest information gain value is used as the root node.
[0029] Optionally, constructing a training set and a test set based on the data feature set using a K-fold cross validation algorithm includes:
[0030] Dividing the data feature set into K data feature subsets;
[0031] Select one of the data feature subsets without repetition to obtain K reference sample sets;
[0032] Each of the reference sample sets is used as a test set in turn, and the remaining K-1 reference sample sets are used as training sets.
[0033] Optionally, after obtaining the result of whether the license is a suspected zombie license, the method further includes:
[0034] Synchronize the merchant-related information corresponding to the suspected zombie license in the result to a preset database;
[0035] Receive a supervision plan formulated based on the relevant information of the suspected zombie license, notify the preset business department to conduct on-site verification according to the supervision plan, and obtain the verification results;
[0036] The license is marked according to the verification result to obtain a label indicating whether the license is a zombie license, and the label is synchronized to the regulatory department.
[0037] In order to solve the above problems, the present invention also provides a zombie license intelligent identification device, the device comprising:
[0038] A data feature set construction module, used to obtain merchant information within a preset historical time period and construct a data feature set using the merchant information;
[0039] A training set and a test set construction module, used to construct a training set and a test set based on the data feature set using a K-fold cross validation algorithm;
[0040] A positive and negative sample labeling module, used to label normal data and abnormal data in the training set and the test set as positive and negative samples respectively;
[0041] A random forest model training module is used to train a pre-constructed random forest model using the training set and the positive and negative samples to obtain a pre-trained random forest model and weights of each data feature, screen the data features according to the weights to obtain retained data features; reconstruct the pre-trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively train the reconstructed random forest model using the test set until a trained random forest model is obtained;
[0042] The screening module is used to use the trained random forest model to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
[0043] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:
[0044] at least one processor; and,
[0045] a memory communicatively connected to the at least one processor; wherein,
[0046] The memory stores a computer program that can be executed by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned zombie license intelligent identification method.
[0047] In order to solve the above problem, the present invention further provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned zombie license intelligent identification method.
[0048] The embodiment of the present invention obtains merchant information within a preset historical time period and uses the merchant information to construct a data feature set; the number of data features can be reduced and the subsequent data processing speed can be improved; K-fold cross-validation algorithm is used to construct K groups of training sets and test sets, and the normal data and abnormal data in the training sets and the test sets are marked as positive and negative samples respectively, and the random forest model is cross-validated, which is conducive to improving the accuracy of the random forest model; the pre-constructed random forest model is trained using the training set and the positive and negative samples to obtain the random forest model that has been preliminarily trained and the weights of each data feature, and the data features with low weights are deleted, and the random forest model that has been preliminarily trained is reconstructed using the retained data features to obtain a reconstructed random forest model, and the data features with low weights are deleted, which is conducive to improving the accuracy and processing efficiency of the random forest model, and the license of the merchant to be analyzed is screened using the trained random forest model to obtain a conclusion on whether the license is a suspected zombie license. Therefore, the zombie license intelligent identification method, device, electronic device and computer-readable storage medium proposed by the present invention can solve the problem of low efficiency and accuracy in zombie license identification. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 A schematic diagram of a flow chart of a method for intelligently identifying zombie licenses provided by an embodiment of the present invention;
[0050] Figure 2 for Figure 1 A schematic diagram of a detailed implementation process of one of the steps in the method for intelligent identification of zombie licenses shown;
[0051] Figure 3 for Figure 1 A schematic diagram of a detailed implementation flow of another step in the method for intelligently identifying zombie licenses is shown;
[0052] Figure 4 A functional module diagram of a zombie license intelligent identification device provided by an embodiment of the present invention;
[0053] Figure 5 A schematic diagram of the structure of an electronic device for implementing the method for intelligently identifying zombie licenses provided by an embodiment of the present invention.
[0054] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0055] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.
[0056] The embodiment of the present application provides a method for intelligently identifying zombie licenses. The execution subject of the method for intelligently identifying zombie licenses includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for intelligently identifying zombie licenses can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.
[0057] Reference Figure 1 FIG. 1 is a flow chart of a method for intelligently identifying zombie licenses provided by an embodiment of the present invention. In this embodiment, the method for intelligently identifying zombie licenses includes:
[0058] S1. Obtain merchant information within a preset historical time period, and use the merchant information to construct a data feature set.
[0059] In the embodiment of the present invention, the merchant information includes: business information, supervision records, historical credit records of stakeholders, etc. Among them, the business information includes: registration, food and catering license, business activity (social security payment, provident fund payment, tax payment, branch opening, investment capital and public opinion news information), annual report, website, funding and awards, etc.; the supervision record includes: the time of daily / special supervision, inspection results, the use of open kitchens and stoves, complaints and reports, etc.; the historical credit records of stakeholders include: historical credit risks of relevant stakeholders such as legal persons and executors, etc.
[0060] In the embodiment of the present invention, crawler technology can be used to obtain merchant information within a preset historical time period from the database of the website corresponding to the regulatory bureau, the food delivery platform, etc. The preset historical time period can be set according to actual needs, such as three years, five years or ten years.
[0061] In detail, the step of constructing a data feature set using the merchant information in S1 includes:
[0062] Cleaning the merchant information, and performing multi-dimensional splitting on the cleaned merchant information to obtain multi-dimensional data;
[0063] Extracting information features of the plurality of dimensional data and association features between the dimensional data, and constructing a corresponding association relationship table using the information features and the association features;
[0064] The association relationship table is used to construct a data feature set.
[0065] In detail, the embodiment of the present invention uses preset rules to identify dirty data in the merchant information and delete the dirty data in the merchant information. The dirty data includes data that is not within a given range or is meaningless to actual business, illegal data format, and data with non-standard coding and ambiguous business logic. For example, anomalies such as unclear user gender, non-numeric tax amount balance field, and non-date time information can all be judged as dirty data.
[0066] In one embodiment of the present invention, the preset rule may adopt a regular expression.
[0067] Among them, the multi-dimensional splitting is the process of classifying the merchant information. After the merchant information is split into dimensions, dimensional data such as one address with multiple certificates, the number of companies that cannot be contacted due to changes in registered addresses, the activity of food delivery platforms, the number of false agents by agents, abnormal supervision results, and the frequency of changes in registration licenses can be obtained.
[0068] Furthermore, the embodiment of the present invention extracts the correlation features between the dimensional data, constructs a correlation table of certificates, licenses, people, supervision, and equipment, and uses the correlation table to form a data feature set of the subject, personnel, equipment, and behavior of the food and catering enterprise.
[0069] The embodiment of the present invention extracts information such as business information, regulatory records, historical credit records of stakeholders, etc., and comprehensively obtains merchant information to avoid low screening accuracy due to incomplete information.
[0070] S2. Based on the data feature set, a training set and a test set are constructed using a K-fold cross-validation algorithm.
[0071] In the embodiment of the present invention, the K-fold cross-validation is a method of statistically cutting data samples into smaller subsets, which is used to divide the initial sampling data into K sub-samples, a single sub-sample is retained as data for validating the model, and the other K-1 samples are used for training.
[0072] In detail, the S2 includes:
[0073] Dividing the data feature set into K data feature subsets;
[0074] Select one of the data feature subsets without repetition to obtain K reference sample sets;
[0075] Each of the reference sample sets is used as a test set in turn, and the remaining K-1 reference sample sets are used as training sets.
[0076] Preferably, in one embodiment of the present invention, a five-fold cross-validation algorithm can be used, that is, the K value is 5, and the data feature set is evenly divided into 5 parts to obtain 5 reference sample sets. In each cross-validation process, 4 of the reference sample sets are selected to form a training set, and the remaining 1 reference sample set is used as a test set for zombie license recognition effect.
[0077] S3. Mark the normal data and abnormal data in the training set and the test set as positive and negative samples, respectively.
[0078] In detail, S3 includes:
[0079] The relevant data in the training set and the test set that have no historical supervision evidence in the one-address-multiple-certificates are defined as abnormal data and marked as negative samples;
[0080] The relevant data of active merchants in the training set and the test set whose business addresses and business scopes are consistent with the licenses are defined as normal data and marked as positive samples.
[0081] S4. Use the training set and the positive and negative samples to train the pre-constructed random forest model to obtain a preliminarily trained random forest model and weights of each data feature, and screen the data features according to the weights to obtain retained data features.
[0082] In the embodiment of the present invention, the random forest model (RF) is a model that integrates multiple trees through the idea of ensemble learning, and its basic unit is a decision tree. Taking the classification problem as an example, each decision tree is a classifier. For an input sample, N trees will have N classification results, and the random forest integrates all classification voting results, and designates the category with the most votes as the final output, thereby obtaining the optimal category.
[0083] For details, see Figure 2 As shown, the S4 includes:
[0084] S41, using the decision layer in the pre-built random forest model to select a root node from the data features in the training set;
[0085] S42, according to the root node, using the classification layer in the pre-built random forest model to classify the remaining data features in the training set, to obtain a plurality of leaf nodes and a random forest model that has been preliminarily trained;
[0086] S43, according to the positive and negative samples, using the activation function in the pre-built random forest model to calculate the Gini coefficient of each leaf node, and obtain the weight of each data feature;
[0087] S44, judging whether the weight meets a preset condition;
[0088] If the weight does not meet the preset condition, execute S45, adjust the parameters of the random forest model that has been preliminarily trained, delete the data features corresponding to the weight that does not meet the preset condition, and return to the step S42;
[0089] If the weights meet the preset conditions, then S46, the retained data features are obtained.
[0090] In one embodiment of the present invention, the Gini coefficient can be obtained by calculating the loss between the leaf node and its corresponding positive and negative samples, and the Gini coefficient is used as the weight of each data feature to evaluate whether the feature of each leaf node in the pre-constructed random forest is the optimal feature. Therefore, the preset condition can be set to whether the weight is less than the preset coefficient. If the weight is not less than the preset coefficient, it means that the weight does not meet the preset condition. If the weight is less than the preset coefficient, it means that the weight meets the preset condition. The preset coefficient can be set to 0.1, or it can be set according to the actual business scenario. The parameters of the random forest model that has been preliminarily trained include the number of data features and the depth of the decision tree.
[0091] For details, see Figure 3 As shown, the S41 includes:
[0092] S411, using the decision algorithm in the decision layer to calculate the information gain value of the data feature of the training data in the merchant information;
[0093] Wherein, the decision algorithm includes:
[0094]
[0095] Among them, C i The information gain value of the data feature, E i represents the expected feature vector of the i-th data feature, The feature vector covariance representing the feature expectation of the i-th data feature, trace() represents the spatial filtering function.
[0096] S412: Use the data feature with the highest information gain value as the root node.
[0097] The implementation of the present invention deletes data features corresponding to weights that do not meet preset conditions, thereby improving the accuracy of the random forest model.
[0098] S5. Reconstruct the random forest model that has been preliminarily trained using the retained data features to obtain a reconstructed random forest model, and iteratively train the reconstructed random forest model using the test set until a trained random forest model is obtained. In an embodiment of the present invention, the pre-constructed random forest model is trained using the training set and the positive and negative samples to obtain a random forest model that has been preliminarily trained and the weights of each data feature, and the random forest model is reconstructed using the weights to improve the detection accuracy of the random forest model.
[0099] In an embodiment of the present invention, the reconstructed random forest model is tested using the test set. When the test fails, the reconstructed random forest model is trained again until the test passes, thereby obtaining a trained random forest model.
[0100] The embodiment of the present invention uses a K-fold cross-validation algorithm to construct K groups of training sets and test sets to cross-validate the random forest model, which is beneficial to improving the accuracy of the random forest model, thereby improving the accuracy of zombie license screening.
[0101] S6. Using the trained random forest model, screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
[0102] In an embodiment of the present invention, data features of the merchant information to be analyzed are extracted, the data features of the merchant information to be analyzed are input into the trained random forest model, the license of the merchant to be analyzed is judged, and a conclusion is obtained as to whether the license is a suspected zombie license.
[0103] In detail, after S5, the method further includes:
[0104] Synchronize the merchant-related information corresponding to the suspected zombie license in the result to a preset database;
[0105] Receive a supervision plan formulated based on the relevant information of the suspected zombie license, notify the preset business department to conduct on-site verification according to the supervision plan, obtain the verification result, mark the license according to the verification result, obtain a label of whether the license is a zombie license, and synchronize the label to the supervision department.
[0106] In one embodiment of the present invention, when it is concluded that the license of the merchant to be analyzed is a suspected zombie license, the merchant information corresponding to the suspected zombie license is synchronized to the credit file, and also synchronized to the preset subject library. The relevant business department specifies the supervision plan and dispatches corresponding staff to conduct on-site verification. When the verification result is: the business is not operating at the address on the license, the suspected zombie license is marked as a zombie license, and the merchant corresponding to the suspected zombie license is marked as "no evidence", and synchronized to the corresponding supervision department.
[0107] In another embodiment of the present invention, the address, company list and legal person credit risk of the merchant corresponding to the suspected zombie license can also be provided to the new license acceptance personnel. The new license acceptance personnel can conduct on-site verification of the merchant and require relevant personnel of the merchant to provide legal use rights materials. After the on-site verification, the new license acceptance personnel will mark the suspected zombie license and simultaneously send it to the corresponding regulatory department.
[0108] The embodiment of the present invention obtains merchant information within a preset historical time period and uses the merchant information to construct a data feature set; the number of data features can be reduced and the subsequent data processing speed can be improved; K-fold cross-validation algorithm is used to construct K groups of training sets and test sets, and the normal data and abnormal data in the training set and the test set are marked as positive and negative samples respectively, and the random forest model is cross-validated, which is conducive to improving the accuracy of the random forest model; the pre-constructed random forest model is trained using the training set and the positive and negative samples to obtain the random forest model that has been preliminarily trained and the weights of each data feature, and the data features with low weights are deleted, and the random forest model that has been preliminarily trained is reconstructed using the retained data features to obtain a reconstructed random forest model, and the data features with low weights are deleted, which is conducive to improving the accuracy and processing efficiency of the random forest model, and the license of the merchant to be analyzed is screened using the trained random forest model to obtain a conclusion on whether the license is a suspected zombie license. Therefore, the zombie license intelligent identification method proposed by the present invention can solve the problem of low efficiency and accuracy in zombie license identification.
[0109] like Figure 4 , which is a functional module diagram of a zombie license intelligent identification device provided by an embodiment of the present invention.
[0110] The zombie license intelligent identification device 100 of the present invention can be installed in an electronic device. According to the functions to be implemented, the zombie license intelligent identification device 100 can include a data feature set construction module 101, a training set and test set construction module 102, a positive and negative sample labeling module 103, a random forest model training module 104 and a screening module 105. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.
[0111] In this embodiment, the functions of each module / unit are as follows:
[0112] The data feature set construction module 101 is used to obtain merchant information within a preset historical time period and construct a data feature set using the merchant information;
[0113] The training set and test set construction module 102 is used to construct a training set and a test set based on the data feature set using a K-fold cross-validation algorithm;
[0114] The positive and negative sample labeling module 103 is used to label the normal data and abnormal data in the training set and the test set as positive and negative samples respectively;
[0115] The random forest model training module 104 is used to train the pre-constructed random forest model using the training set and the positive and negative samples to obtain a pre-trained random forest model and weights of each data feature, screen the data features according to the weights to obtain retained data features; reconstruct the pre-trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively train the reconstructed random forest model using the test set until a trained random forest model is obtained;
[0116] The screening module 105 is used to screen the license of the merchant to be analyzed by using the trained random forest model to obtain a result of whether the license is a suspected zombie license.
[0117] In detail, each module described in the zombie license intelligent identification device 100 in the embodiment of the present invention is used in the same manner as described above. Figures 1 to 3 The technical means are the same as the zombie license intelligent identification method described in, and can produce the same technical effects, so they will not be repeated here.
[0118] like Figure 5 FIG. 1 is a schematic diagram of the structure of an electronic device for implementing a method for intelligently identifying zombie licenses according to an embodiment of the present invention.
[0119] The electronic device 1 may include a processor 10, a memory 11, a communication bus 12, and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a zombie license intelligent identification program.
[0120] The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 10 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11 (for example, executing a zombie license intelligent identification program, etc.), and calls data stored in the memory 11 to execute various functions of the electronic device and process data.
[0121] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 11 may also include both an internal storage unit of the electronic device and an external storage device. The memory 11 may not only be used to store application software and various types of data installed in the electronic device, such as the code of the zombie license intelligent identification program, but also be used to temporarily store data that has been output or is to be output.
[0122] The communication bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.
[0123] The communication interface 13 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.
[0124] Figure 5 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 5 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.
[0125] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.
[0126] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.
[0127] The zombie license intelligent identification program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve:
[0128] Acquire merchant information within a preset historical time period, and use the merchant information to construct a data feature set;
[0129] Based on the data feature set, a training set and a test set are constructed using a K-fold cross validation algorithm;
[0130] Marking normal data and abnormal data in the training set and the test set as positive and negative samples respectively;
[0131] Using the training set and the positive and negative samples to train a pre-constructed random forest model, obtaining a preliminarily trained random forest model and weights of each data feature, and screening the data features according to the weights to obtain retained data features;
[0132] Reconstructing the initially trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively training the reconstructed random forest model using the test set until a trained random forest model is obtained;
[0133] The trained random forest model is used to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
[0134] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.
[0135] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).
[0136] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:
[0137] Acquire merchant information within a preset historical time period, and use the merchant information to construct a data feature set;
[0138] Based on the data feature set, a training set and a test set are constructed using a K-fold cross validation algorithm;
[0139] Marking normal data and abnormal data in the training set and the test set as positive and negative samples respectively;
[0140] Using the training set and the positive and negative samples to train a pre-constructed random forest model, obtaining a preliminarily trained random forest model and weights of each data feature, and screening the data features according to the weights to obtain retained data features;
[0141] Reconstructing the initially trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively training the reconstructed random forest model using the test set until a trained random forest model is obtained;
[0142] The trained random forest model is used to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
[0143] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.
[0144] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0145] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.
[0146] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0147] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.
[0148] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.
[0149] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0150] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.
[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.
Claims
1. A method for intelligently identifying zombie licenses, characterized in that: The method comprises: Obtain merchant information within a preset historical time period, identify and delete dirty data in the merchant information using preset rules, and construct a data feature set using the merchant information after deleting the dirty data. The dirty data includes data that is not within a given range or is meaningless to actual business, has an illegal data format, and has non-standard coding and ambiguous business logic; Based on the data feature set, a training set and a test set are constructed using a K-fold cross validation algorithm; The relevant data in the training set and the test set with multiple licenses for one address that have no historical supervision are defined as abnormal data and marked as negative samples, and the relevant data in the training set and the test set with active merchants whose business addresses and business scopes are consistent with the licenses are defined as normal data and marked as positive samples; Using the training set and the positive samples and the negative samples to train a pre-constructed random forest model, obtaining a preliminarily trained random forest model and weights of each data feature, and screening the data features according to the weights to obtain retained data features; Reconstructing the initially trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively training the reconstructed random forest model using the test set until a trained random forest model is obtained; The trained random forest model is used to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
2. The method for intelligently identifying zombie licenses as claimed in claim 1, characterized in that: The step of constructing a data feature set using the merchant information includes: Cleaning the merchant information, and performing multi-dimensional splitting on the cleaned merchant information to obtain multi-dimensional data; Extracting information features of the plurality of dimensional data and association features between the dimensional data, and constructing a corresponding association relationship table using the information features and the association features; The association relationship table is used to construct a data feature set.
3. The method for intelligently identifying zombie licenses as claimed in claim 1, characterized in that: The pre-constructed random forest model is trained by using the training set and the positive samples and the negative samples to obtain a preliminarily trained random forest model and weights of each data feature, and the data features are screened according to the weights to obtain the retained data features, including: Using the decision layer in the pre-built random forest model, a root node is selected from the data features in the training set; According to the root node, the classification layer in the pre-built random forest model is used to classify the remaining data features in the training set to obtain a plurality of leaf nodes and a random forest model that has been preliminarily trained; According to the average accuracy of the positive samples and the negative samples, the Gini coefficient of each leaf node is calculated using the activation function in the pre-built random forest model to obtain the weight of each data feature; Determining whether the weight meets a preset condition; If the weight does not meet the preset condition, adjust the parameters of the random forest model that has been preliminarily trained, delete the data features corresponding to the weights that do not meet the preset condition, and return to execute the step of classifying the remaining data features in the training set according to the root node using the classification layer in the pre-constructed random forest model; If the weights meet the preset conditions, the retained data features are obtained.
4. The method for intelligently identifying zombie licenses as claimed in claim 3, characterized in that: The step of using the decision layer in the pre-built random forest model to select a root node from the data features in the training set includes: Calculating the information gain value of the data feature of the training data in the merchant information using the decision algorithm in the decision layer; The data feature with the highest information gain value is used as the root node.
5. The method for intelligently identifying zombie licenses as claimed in claim 1, characterized in that: The method of constructing a training set and a test set based on the data feature set using a K-fold cross validation algorithm includes: Dividing the data feature set into K data feature subsets; Select one of the data feature subsets without repetition to obtain K reference sample sets; Each of the reference sample sets is used as a test set in turn, and the remaining K-1 reference sample sets are used as training sets.
6. The method for intelligently identifying zombie licenses according to any one of claims 1 to 5, characterized in that: After obtaining the result of whether the license is a suspected zombie license, the method further includes: Synchronize the merchant-related information corresponding to the suspected zombie license in the result to a preset database; Receive a supervision plan formulated based on the relevant information of the suspected zombie license, notify the preset business department to conduct on-site verification according to the supervision plan, and obtain the verification results; The license is marked according to the verification result to obtain a label indicating whether the license is a zombie license, and the label is synchronized to the regulatory department.
7. A zombie license intelligent identification device, characterized in that: The device comprises: A data feature set construction module is used to obtain merchant information within a preset historical time period, identify and delete dirty data in the merchant information using preset rules, and construct a data feature set using the merchant information after deleting the dirty data. The dirty data includes data that is not within a given range or is meaningless to actual business, has an illegal data format, and has non-standard coding and ambiguous business logic; A training set and a test set construction module, used to construct a training set and a test set based on the data feature set using a K-fold cross validation algorithm; A positive and negative sample labeling module, used to define the relevant data in the training set and the test set that are not verified by historical supervision as abnormal data and label them as negative samples, and define the relevant data of active merchants in the training set and the test set whose business address and business scope are consistent with the license as normal data and label them as positive samples; A random forest model training module is used to train a pre-constructed random forest model using the training set and the positive samples and the negative samples to obtain a preliminarily trained random forest model and weights of each data feature, screen the data features according to the weights to obtain retained data features; reconstruct the preliminarily trained random forest model using the retained data features to obtain a reconstructed random forest model, and iteratively train the reconstructed random forest model using the test set until a trained random forest model is obtained; The screening module is used to use the trained random forest model to screen the license of the merchant to be analyzed to obtain a result of whether the license is a suspected zombie license.
8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the zombie license intelligent identification method according to any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for intelligently identifying zombie licenses according to any one of claims 1 to 6 is implemented.
Citation Information
Patent Citations
An over-range operation judgment method for online catering merchants
CN109934596A
Order abnormality intelligent detection method and device, electronic equipment and storage medium
CN113688923A
