Iot card identification method and device

Abstract

The application provides an Internet of Things card identification method and device, relates to the technical field of Internet of Things, and solves the problem of low identification accuracy of abnormal Internet of Things cards. The method comprises the following steps: receiving first sample data from an operator node; the first sample data is obtained according to M groups of second sample data; the kth group of second sample data in the M groups of second sample data is obtained according to third sample data and the kth group of fourth sample data; the M enterprise nodes belong to the same industry; k and M are positive integers, and 1 < k ≤ M; receiving M groups of gradient information from the M enterprise nodes; the kth group of gradient information in the M groups of gradient information is obtained by training a kth enterprise local identification model according to the kth group of second sample data; sending an industry global identification model to the M enterprise nodes; the industry global identification model is used for identifying Internet of Things cards; and the industry global identification model is obtained according to at least one of the following: the first sample data or the M groups of gradient information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, and in particular to an Internet of Things card identification method and device. BACKGROUND

[0002] In the 5G Internet of Everything era, operators vigorously develop Internet of Things services for enterprise customers (to B), and the operators will sell Internet of Things cards to various enterprises, and the enterprises will sell products or services along with the Internet of Things cards to users. Due to the special nature of the Internet of Things card without real-name authentication, abnormal use behaviors such as misappropriation, transfer and abuse of the Internet of Things card often occur, causing great economic losses to enterprises and operators.

[0003] At present, the abnormal identification of the Internet of Things card only uses the historical data recorded by the operators or the enterprises to construct identification rules or identification models. However, the user data of the enterprise side is small, and the user data of different enterprises in the same industry cannot be shared, so the data is not comprehensive, which will result in low accuracy of identifying abnormal Internet of Things cards. In addition, although the data of the operator side is large, it lacks core business data of enterprise users, which will also result in low accuracy of identifying abnormal Internet of Things cards, so that the abnormal Internet of Things cards cannot be effectively detected. SUMMARY

[0004] The present application provides an Internet of Things card identification method and device, which solves the problem of low accuracy of identifying abnormal Internet of Things cards.

[0005] To achieve the above purpose, the present application adopts the following technical scheme:

[0006] In a first aspect, an Internet of Things card identification method is provided. The method comprises: receiving first sample data from an operator node. The first sample data is obtained according to M groups of second sample data; the kth group of second sample data in the M groups of second sample data is obtained according to third sample data and the kth group of fourth sample data; the third sample data is data corresponding to M enterprise nodes in the user data recorded by the operator node, and the kth group of fourth sample data is user data of the kth enterprise node in the M enterprise nodes; the M enterprise nodes belong to the same industry; k and M are positive integers, and 1

[0007] Further, the first sample data is obtained according to an industry common feature extracted from the M groups of second sample data.

[0008] Further, the kth group of second sample data is obtained according to data of the same user in the third sample data and the kth group of fourth sample data.

[0009] In a possible design, the method of the first aspect further includes: sending a global parameter to the M enterprise nodes. The global parameter is obtained according to the first sample data training an industry global identification model. Receiving M groups of gradient information from the M enterprise nodes. The kth group of gradient information in the M groups of gradient information is obtained according to the global parameter and the kth group of second sample data training the kth enterprise local identification model. Updating the global parameter according to the M groups of gradient information. If the accuracy of the industry global identification model determined by the updated global parameter is greater than or equal to an accuracy threshold, then determining the industry global identification model according to the updated global parameter.

[0010] The second aspect provides an Internet of Things card identification method. The method includes: receiving an industry global identification model from a service node. The industry global identification model is used to identify an Internet of Things card. The industry global identification model is obtained according to at least one of the following: first sample data or M groups of gradient information. The first sample data is obtained according to M groups of second sample data. The kth group of second sample data in the M groups of second sample data is obtained according to third sample data and the kth group of fourth sample data. The third sample data is data corresponding to the M enterprise nodes in user data recorded by an operator node. The kth group of fourth sample data is data of the kth enterprise node in the M enterprise nodes. The M enterprise nodes belong to the same industry. k and M are positive integers, and 1 < k ≤ M. The kth group of gradient information in the M groups of gradient information is obtained according to the kth group of second sample data training the kth enterprise local identification model. The kth enterprise local identification model corresponds to the kth enterprise node. Training the industry global identification model according to enterprise individual features to obtain an enterprise individualized identification model.

[0011] In a third aspect, an Internet of Things card identification apparatus is provided. The apparatus comprises a receiving module and a sending module. The receiving module is configured to receive first sample data from an operator node. The first sample data is obtained based on M groups of second sample data. The kth group of second sample data in the M groups of second sample data is obtained based on third sample data and the kth group of fourth sample data. The third sample data is data corresponding to M enterprise nodes in user data recorded by the operator node, and the kth group of fourth sample data is user data of a kth enterprise node in the M enterprise nodes. The M enterprise nodes belong to the same industry. k and M are positive integers, and 1 < k ≤ M. The receiving module is further configured to receive M groups of gradient information from the M enterprise nodes. The kth group of gradient information in the M groups of gradient information is obtained based on training of a kth enterprise local identification model by the kth group of second sample data. The kth enterprise local identification model corresponds to the kth enterprise node. The sending module is configured to send an industry global identification model to the M enterprise nodes. The industry global identification model is used to identify Internet of Things cards. The industry global identification model is obtained based on at least one of the first sample data or the M groups of gradient information.

[0012] Further, the first sample data is obtained based on industry common features extracted from the M groups of second sample data.

[0013] Further, the kth group of second sample data is obtained based on data of the same user in the third sample data and the kth group of fourth sample data.

[0014] In a possible design, the apparatus of the third aspect further comprises a processing module. The sending module is configured to send a global parameter to the M enterprise nodes. The global parameter is obtained based on training of the industry global identification model by the first sample data. The receiving module is configured to receive M groups of gradient information from the M enterprise nodes. The kth group of gradient information in the M groups of gradient information is obtained based on training of the kth enterprise local identification model by the global parameter and the kth group of second sample data. The processing module is configured to update the global parameter based on the M groups of gradient information. If the accuracy of the industry global identification model determined by the updated global parameter is greater than or equal to an accuracy threshold, the industry global identification model is determined based on the updated global parameter.

[0015] Optionally, the receiving module and the sending module can be integrated into a transceiver module. The transceiver module is configured to implement the receiving and sending functions of the Internet of Things card identification apparatus of the third aspect.

[0016] In a fourth aspect, an Internet of Things card identification apparatus is provided. The apparatus comprises a receiving module and a processing module. The receiving module is configured to receive an industry global identification model from a service node. The industry global identification model is configured to identify an Internet of Things card. The industry global identification model is obtained according to at least one of the following: first sample data or M sets of gradient information. The first sample data is obtained according to M sets of second sample data. The kth set of second sample data in the M sets of second sample data is obtained according to third sample data and a kth set of fourth sample data. The third sample data is data corresponding to M enterprise nodes in user data recorded by an operator node. The kth set of fourth sample data is user data of a kth enterprise node in the M enterprise nodes. The M enterprise nodes belong to the same industry. k and M are positive integers, and 1 < k ≤ M. The kth set of gradient information in the M sets of gradient information is obtained by training a kth enterprise local identification model according to the kth set of second sample data. The kth enterprise local identification model corresponds to the kth enterprise node. The processing module is configured to train the industry global identification model according to enterprise individual characteristics to obtain an enterprise individualized identification model.

[0017] Optionally, the Internet of Things card identification apparatus of the fourth aspect can further comprise a sending module. The sending module is configured to implement the sending function of the Internet of Things card identification apparatus of the fourth aspect.

[0018] Optionally, the receiving module and the sending module can be integrated into a transceiving module. The transceiving module is configured to implement the transceiving function of the Internet of Things card identification apparatus of the fourth aspect.

[0019] In a fifth aspect, an Internet of Things card identification apparatus is provided. The apparatus comprises a memory and a processor. The memory is configured to store computer execution instructions. The processor is connected to the memory through a bus. When the apparatus is running, the processor executes the computer execution instructions stored in the memory, so that the apparatus performs the Internet of Things card identification method of any one of the first aspect to the second aspect.

[0020] In a sixth aspect, a computer readable storage medium is provided. The computer readable storage medium stores computer instructions. When the computer instructions are run on a computer, the computer performs the Internet of Things card identification method of any one of the first aspect to the second aspect.

[0021] It should be noted that the above computer instructions can be stored on a computer storage medium in whole or in part. The computer storage medium can be packaged together with the processor of the Internet of Things card identification apparatus, or can be packaged separately from the processor of the Internet of Things card identification apparatus. The embodiments of the present application do not limit this.

[0022] In the present application, the names of the above-mentioned Internet of Things card identification apparatus do not constitute a limitation on the devices or functional modules themselves, and in actual implementation, these devices or functional modules can appear under other names. As long as the functions of each device or functional module are similar to those of the present application, they belong to the scope of the claims of the present application and their equivalents.

[0023] Based on the first aspect to the sixth aspect, based on the fusion of the sample data of the operator and the sample data of different enterprises in the same industry, the second sample data is constructed, the industry common characteristics in the second sample data are extracted to construct the first sample data, and then the industry global identification model of the same Internet of Things industry is trained using the first sample data, or the gradient information obtained by training the enterprise local identification model using the second sample data constructed based on the fusion of the sample data of the operator and the sample data of different enterprises in the same industry, and the industry global identification model is obtained according to the gradient information. The data of different enterprises in the same industry and the data of the operator are combined to construct the industry global identification model, the sample data characteristics are diversified, and the accuracy of the Internet of Things card anomaly identification can be improved.

[0024] Further, the enterprise personalized identification model can be customized by combining the industry global identification model and the enterprise individual characteristics, so as to further improve the accuracy of the enterprise in identifying the abnormal Internet of Things card.

[0025] These aspects or other aspects of the present application will be more apparent in the following description. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 An architecture schematic diagram of an Internet of Things card identification system provided by an embodiment of the present application;

[0027] Figure 2 A flowchart of an Internet of Things card identification method provided by an embodiment of the present application;

[0028] Figure 3 A flowchart of another Internet of Things card identification method provided by an embodiment of the present application;

[0029] Figure 4 A structure schematic diagram of an Internet of Things card identification apparatus provided by an embodiment of the present application;

[0030] Figure 5 A structure schematic diagram of another Internet of Things card identification apparatus provided by an embodiment of the present application;

[0031] Figure 6 A structure schematic diagram of another Internet of Things card identification apparatus provided by an embodiment of the present application; DETAILED DESCRIPTION

[0032] With reference to the drawings and the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0033] It should be noted that in the embodiments of the present application, the words such as "exemplary" or "for example" are used to mean serving as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or having more advantages than other embodiments or design schemes. Rather, the words such as "exemplary" or "for example" are used in the specific manner to present the relevant concept.

[0034] In order to clearly describe the technical solutions in the embodiments of the present application, in the embodiments of the present application, the words such as "first", "second", and the like are used to distinguish the same or similar items or items with basically the same function and role. A person skilled in the art can understand that the words such as "first", "second", and the like are not used to limit the quantity and execution order.

[0035] The Internet of Things card identification method in the embodiments of the present application can be applied to Figure 1 The Internet of Things card identification system shown. The Internet of Things card identification system can include a service node, an operator node, and M enterprise nodes. The service node is in communication connection with the operator node and the M enterprise nodes, respectively. The operator node is in communication connection with the M enterprise nodes, respectively. The M enterprise nodes belong to the same industry, M is a positive integer, and M>1, for example, Tesla enterprises, BYD enterprises, and the like in the Internet of Vehicles industry.

[0036] It should be noted that the service node, the operator node, and the enterprise node described above can be a device, a chip or a chip system in the device, or a system or a network containing the device, and the embodiments of the present application are not limited thereto.

[0037] Optionally, the service node, the operator node, and the enterprise node can be a server, a notebook computer, a computer, or the like, or can be a virtual machine (VM) deployed on a physical machine, without specific limitation here.

[0038] The operator node and the M enterprise nodes can all construct sample data based on the user data, the sample data of each node can be distinguished using a user identity document (ID), which can be an Internet of Things card number, that is, the telephone number of the user to whom the Internet of Things card belongs. The operator node and the enterprise node can align the encrypted samples through the user ID, which can use a homomophic encryption algorithm, including an RSA encryption algorithm, a Pallier encryption algorithm, etc., which can be selected according to actual application scenarios and data processing requirements.

[0039] The sample data of the operator node can include network side feature data and enterprise side feature data. The network side feature data can include user behavior data, user call data, user location data, user network signaling data, user resident area, user service frequency, service daily traffic, mobile state information, etc., and the enterprise side feature data can include package type, consumption amount information, home location information, etc.

[0040] Specifically, the operator node constructs third sample data based on user data, and the M enterprise nodes respectively construct M groups of fourth sample data based on enterprise user data. The operator node determines the sample data of the same user ID in the fourth sample data corresponding to each enterprise node and the third sample data based on an encryption alignment algorithm, obtains M groups of second sample data, and extracts industry common features of the M groups of second sample data to form first sample data.

[0041] It should be noted that the service node can receive the first sample data from the operator node to construct an industry global recognition model, and / or receive gradient information from the M enterprise nodes to construct an industry global recognition model, the gradient information being obtained by training the enterprise local recognition model according to the second sample data by each enterprise node. The data transmission between any two of the service node, the operator node and the enterprise node adopts encrypted transmission to avoid leaking user privacy.

[0042] Further, the service node sends the industry global recognition model to each enterprise node, each enterprise node trains the industry global recognition model based on the enterprise individual features of each enterprise to form an enterprise individual recognition model suitable for each enterprise, so as to further improve the accuracy of the enterprise in identifying abnormal Internet of Things cards.

[0043] It should be noted that the service node can be a device or a device set including multiple devices, and the multiple devices can be devices of the same type or devices of multiple types. The service node can construct different industry global recognition models for different Internet of Things industries.

[0044] The Internet of Things card identification system formed by the nodes can be used to implement the Internet of Things card identification method described in the method embodiment.

[0045] Figure 2 A flowchart of an Internet of Things card identification method provided by the embodiments of the present application is shown. The method is applicable to the Internet of Things card identification system shown in Figure 1 As shown in Figure 2 The method includes S201-S206.

[0046] S201, the operator node obtains the kth set of second sample data according to the third sample data and the kth set of fourth sample data.

[0047] The third sample data is the data corresponding to the M enterprise nodes in the user data recorded by the operator node, and the kth set of fourth sample data can be the user data of the kth enterprise node in the M enterprise nodes shown in Figure 1 The third sample data and the fourth sample data both take the Internet of Things card number as the user ID. The k is a positive integer, and 1

[0048] Specifically, the operator node performs sample alignment on the kth set of fourth sample data and the fourth sample data of the kth enterprise node based on an encryption alignment algorithm, determines the sample data with the same user ID in the third sample data and the kth set of fourth sample data, integrates the features in the third sample data and the fourth sample data with the same user ID, and obtains the kth set of second sample data. Using the encryption alignment algorithm to fuse the data on both sides can prevent private data information from being leaked.

[0049] For example, the operator node can use the Pallier encryption algorithm to generate a key pair, send the public key in the key pair to the kth enterprise node for encrypting the kth set of fourth sample data, and receive the kth set of fourth sample data after encryption. The operator node determines the sample data with the same user ID in the third sample data and the kth set of fourth sample data according to the user ID, decrypts the fourth sample data according to the private key in the generated key pair, and integrates the features of the third sample data and the fourth sample data with the same user ID to obtain the kth set of second sample data.

[0050] For example, the third sample data with the same user ID is (a, b, c, d), the fourth sample data is (a, c, e, f, g), and the second sample data of the user ID after integration is (a, b, c, d, e, f, g). In other words, the second sample data of the user ID integrates all features in the sample data of the user ID recorded by the operator node and the kth enterprise node, and the sample data of the user recorded by the operator node and the sample data of the user recorded by the enterprise node are complementary in features to form complete second sample data.

[0051] In a possible design, the service node can also generate a plurality of pairs of keys based on the encryption algorithm, and distribute the key pairs to the operator node and the M enterprise nodes. The plurality of pairs of keys can be the same or different, and the embodiments of the present application are not limited thereto.

[0052] Optionally, the operator node can send the generated key pair to the service node, or send the key generation parameter to the service node, and the service node can calculate the key from the key generation parameter by using the same key generation algorithm, so as to further improve the security.

[0053] It can be understood that the encryption alignment algorithm is completed by interaction between the operator node and the enterprise node, and therefore, the kth enterprise node can also obtain the kth set of second sample data based on the encryption alignment algorithm.

[0054] S202, the operator node obtains first sample data according to the M sets of second sample data.

[0055] Specifically, the operator node obtains the M sets of second sample data by the above step S201 with the M enterprise nodes respectively, and then extracts the industry common features of the M enterprises from the M sets of second sample data to form the first sample data.

[0056] Exemplarily, three enterprise nodes in the public utility industry, the second sample data of a user with an ID of 1111 in the enterprise node 1 is (a1, b1, c1, d1, e1, f1, g1), the second sample data of a user with an ID of 2222 in the enterprise node 2 is (a2, b2, c2, d2, e2, h2, j2, k2), and the second sample data of a user with an ID of 3333 in the enterprise node 3 is (a3, b3, c3, d3, e3, l3, m3), and the industry common characteristics of the three are (a, b, c, d, e), and the first sample data of the user with the ID of 1111 is (a1, b1, c1, d1, e1), the first sample data of the user with the ID of 2222 is (a2, b2, c2, d2, e2), and the first sample data of the user with the ID of 3333 is (a3, b3, c3, d3, e3) based on the industry common characteristics. The industry common characteristics (a, b, c, d, e) of the public utility industry can be user basic attribute characteristics (such as province, city, industry category, membership level, network access time length, terminal type, etc.), service characteristics (such as common services, service preferences, etc.), geographic characteristics (such as permanent residence area, etc.), and the like, which are not limited by the embodiments of the present application.

[0057] S203, the operator node sends the first sample data to the service node.

[0058] Specifically, the operator node encrypts the first sample data and sends it to the service node, and the service node can decrypt the encrypted first sample data based on the key pair sent by the operator node to obtain the first sample data.

[0059] It is worth noting that the keys used for transmission between the operator node and the service node, the keys used for transmission between the operator node and the enterprise node, and the keys used for transmission between the service node and the enterprise node can be the same or different, which is not limited herein.

[0060] S204, the service node obtains an industry global identification model according to the first sample data.

[0061] Specifically, the service node can iteratively train the industry global identification model according to the first sample data until a training end condition is met to obtain a trained industry global identification model. The training end condition can be that the accuracy of the industry global identification model is greater than or equal to a set first accuracy threshold, or that a set maximum number of iterative training is reached.

[0062] S205, the service node sends the industry global identification model to the kth enterprise node.

[0063] Specifically, the service node sends the trained industry global identification model to each enterprise node, and each enterprise node can perform abnormal Internet of Things card identification based on the industry global identification model.

[0064] Further, in S206, the kth enterprise node trains the industry global identification model according to the enterprise individual characteristics to obtain an enterprise individualized identification model.

[0065] Specifically, since the same industry different enterprises have both industry common characteristics and enterprise individual characteristics, each enterprise node can customize an individualized identification model suitable for each enterprise, that is, the kth enterprise node further trains the industry global identification model according to the enterprise individual characteristics of the enterprise node to obtain an enterprise individualized identification model, so that the Internet of Things card identification accuracy of each enterprise is further improved.

[0066] Figure 3 Another flowchart of an Internet of Things card identification method provided by the embodiments of the present application is shown. The method is applicable to the Internet of Things card identification system shown. Figure 1 As shown in the Internet of Things card identification system shown, the method comprises S301-S306. Figure 3

[0067] S301, the kth enterprise node obtains the kth set of second sample data according to the third sample data and the kth set of fourth sample data.

[0068] Specifically, the kth enterprise node in the M enterprise nodes obtains the third sample data from the operator node based on the encryption alignment algorithm, aligns the kth set of fourth sample data with the third sample data, determines the sample data with the same user ID in the third sample data and the kth set of fourth sample data, integrates the features of the third sample data and the kth set of fourth sample data with the same user ID together, and obtains the kth set of second sample data.

[0069] The third sample data obtained by the kth enterprise node from the operator node can be data related only to the kth enterprise node, for example, user data registered to purchase Internet of Things card business in the kth enterprise node.

[0070] Exemplarily, the kth enterprise node in the M enterprise nodes can generate a key pair by using the Pallier encryption algorithm, send the public key in the key pair to the operator node for encrypting the third sample data, receive the encrypted third sample data from the operator node, determine the sample data with the same user ID in the third sample data and the kth set of fourth sample data according to the user ID, decrypt the determined third sample data according to the private key in the generated key pair, and integrate the features of the third sample data and the fourth sample data with the same user ID together to obtain the kth set of second sample data. ​

[0071] Optionally, the kth set of second sample data can also be sent to the kth enterprise node by the operator node according to step S201.

[0072] S302, the kth enterprise node trains the enterprise local recognition model according to the kth set of second sample data, and obtains the kth set of gradient information.

[0073] Specifically, each enterprise node (the kth enterprise node) in the M enterprise nodes obtains a set of second sample data based on the encryption alignment algorithm in step S301, each enterprise node iteratively trains the enterprise local recognition model according to the corresponding obtained second sample data, obtains a set of gradient information, and then sends the trained gradient information to the service node after encryption. The M enterprise nodes correspond to M sets of gradient information. The enterprise local recognition model can be an initialized industry global recognition model distributed by the service node to each enterprise node, or an industry global recognition model obtained in the last round of training, which is not limited by the embodiments of the present application.

[0074] S303, the kth enterprise node sends the kth set of gradient information to the service node.

[0075] S304, the service node obtains the industry global recognition model according to the M sets of gradient information.

[0076] Specifically, the service node decrypts the M sets of gradient information and performs weighted calculation to obtain the global parameter, which is used as the parameter of the industry global recognition model. The M sets of gradient information are obtained by the M enterprise nodes through step S302.

[0077] Further, based on the global parameter of the industry global recognition model obtained from the M sets of gradient information, the service node can optimize and train the industry global recognition model with the global parameter, adjust the global parameter, and thus obtain an industry global recognition model with higher accuracy.

[0078] S305, the service node sends the industry global recognition model to the kth enterprise node.

[0079] S306, the kth enterprise node trains the industry global recognition model according to the enterprise individual characteristics, and obtains the enterprise individualized recognition model.

[0080] Specifically, steps S305-S306 can refer to steps S205-S206 described above, which will not be repeated here.

[0081] It can be understood that the industry global recognition model can be obtained based on steps S201-S204 described above, or based on S301-S304, which is not limited by the embodiments of the present application.

[0082] In a possible design, to further improve the accuracy of the industry global recognition model, the service node can jointly train the industry global recognition model with the M enterprise nodes. In other words, the service node can train the industry global recognition model in combination with the first sample data and the M sets of gradient information. Exemplarily, the training specifically includes the following steps:

[0083] Step 1: The service node sends the global parameter to the M enterprise nodes.

[0084] The global parameter can be obtained by training the industry global recognition model according to the first sample data in the step S204, and the global parameter is sent to the enterprise nodes after being encrypted.

[0085] Step 2: The kth enterprise node trains the kth enterprise local recognition model according to the global parameter and the kth set of second sample data, to obtain the kth set of gradient information.

[0086] Specifically, the kth enterprise node of the M enterprises decrypts the global parameter as the initial parameter for training the enterprise local recognition model, and inputs the kth set of second sample data into the enterprise local recognition model for iterative training, to obtain the kth set of gradient information.

[0087] Step 3: The M enterprise nodes send the M sets of gradient information to the service node.

[0088] Specifically, the M enterprise nodes send the M sets of gradient information obtained through the step 2 to the service node after being encrypted.

[0089] Step 4: The service node updates the global parameter according to the M sets of gradient information.

[0090] Specifically, the service node decrypts the M sets of gradient information, and then performs weighted average calculation on the M sets of gradient information, to determine the weighted average value as the global parameter, and complete the update of the global parameter.

[0091] Exemplarily, the weighted average calculation can refer to the following formula:

[0092]

[0093] wherein, F is the global parameter, F k is the gradient information of the kth enterprise node, M is the number of enterprise nodes, n is the total number of second sample data of the M enterprise nodes, n k is the number of second sample data of the kth enterprise node.

[0094] That is, the service node updates the global parameter in combination with the proportion of the second sample data provided by each enterprise node.

[0095] Step 5, the service node determines whether the accuracy of the industry global recognition model determined by the updated global parameters is greater than or equal to a second accuracy threshold.

[0096] Specifically, if the accuracy of the industry global recognition model determined by the updated global parameters is greater than or equal to the second accuracy threshold, the industry global recognition model is determined according to the updated global parameters.

[0097] Or, if the accuracy of the industry global recognition model determined by the updated global parameters is less than the second accuracy threshold, the updated global parameters are sent to the M enterprise nodes for re-updating training, that is, the above steps 1 to 5 are repeated, and the global parameters in step 1 are the global parameters updated last time.

[0098] Wherein, the second accuracy threshold is different from the first accuracy threshold in the above step S204. It can be understood that the second accuracy threshold is usually greater than the first accuracy threshold.

[0099] In the process of iterative updating training of the above steps 1 to 5, in addition to the encrypted interaction of gradient information and global parameters between the service node and each enterprise node, each enterprise node and the service node independently train the local data set (first sample data or second sample data), and cannot obtain the data of other enterprise nodes, so as to avoid data privacy leakage.

[0100] In another possible design scheme, the service node can obtain the industry global recognition model according to the gradient information reported by a part of enterprise nodes and the second sample data training reported by another part of enterprise nodes.

[0101] In addition, in combination with Figure 2 and Figure 3 It can be known that the operator node and / or the enterprise node can also send M sets of second sample data to the service node, and then the service node extracts the industry common features of the M sets of second sample data to form the first sample data, and trains the industry global recognition model based on the first sample data.

[0102] The above is based on Figure 2 or Figure 3The shown Internet of Things card identification method is based on fusion of sample data of an operator and sample data of different enterprises in the same industry, constructs second sample data, extracts industry common features in the second sample data to construct first sample data, and then uses the first sample data to train an industry global identification model of the same Internet of Things industry, or uses gradient information obtained by training an enterprise local identification model based on the second sample data constructed by fusion of sample data of an operator and sample data of different enterprises in the same industry to obtain the industry global identification model. The data of different enterprises in the same industry and the data of the operator are combined to construct the industry global identification model, the sample data features are diversified, and the accuracy of Internet of Things card anomaly identification can be improved.

[0103] Further, the industry global identification model and the enterprise individual feature can be combined to customize an enterprise individual identification model suitable for the enterprise to further improve the accuracy of identifying abnormal Internet of Things cards.

[0104] The above mainly introduces the scheme provided by the embodiments of the present application from the perspective of the method. To implement the above functions, it contains the hardware structure and / or software module corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments disclosed herein, the embodiments of the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed by hardware or computer software driven hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of the present application.

[0105] The embodiments of the present application can divide the function modules of the Internet of Things card identification device according to the above method examples. For example, each function module can be divided according to each function, or two or more functions can be integrated in one processing module. The above integrated module can be realized in the form of hardware or software function module. Optionally, the division of modules in the embodiments of the present application is illustrative, and is only a logical function division. When actually implemented, there can be another division method.

[0106] Figure 4 A structure diagram of an Internet of Things card identification device provided by the embodiments of the present application is shown. The device is applicable to Figure 1 The shown Internet of Things card identification system. As Figure 4 The Internet of Things card identification device 400 includes a receiving module 401 and a sending module 402.

[0107] The receiving module 401 is configured to receive first sample data from an operator node.

[0108] The first sample data is obtained according to M groups of second sample data; kth group of second sample data in the M groups of second sample data is obtained according to third sample data and kth group of fourth sample data; the third sample data is data corresponding to the M enterprise nodes in user data recorded by the operator node, and the kth group of fourth sample data is user data of a kth enterprise node in the M enterprise nodes; the M enterprise nodes belong to the same industry; k and M are positive integers, and 1 < k ≤ M.

[0109] The receiving module 401 is further configured to receive M groups of gradient information from the M enterprise nodes.

[0110] The kth group of gradient information in the M groups of gradient information is obtained according to the kth enterprise local identification model trained by the kth group of second sample data; and the kth enterprise local identification model corresponds to the kth enterprise node.

[0111] The sending module 402 is configured to send an industry global identification model to the M enterprise nodes. The industry global identification model is used to identify the Internet of Things card; and the industry global identification model is obtained according to at least one of the following: the first sample data or the M groups of gradient information.

[0112] Further, the first sample data is obtained according to industry common features extracted from the M groups of second sample data.

[0113] Further, the kth group of second sample data is obtained according to data of the same user in the third sample data and the kth group of fourth sample data.

[0114] In a possible design, the Internet of Things card identification apparatus 400 further includes a processing module 403.

[0115] The sending module 402 is configured to send a global parameter to the M enterprise nodes. The global parameter is obtained according to the first sample data to train the industry global identification model.

[0116] The receiving module 401 is configured to receive M groups of gradient information from the M enterprise nodes. The kth group of gradient information in the M groups of gradient information is obtained according to the global parameter and the kth group of second sample data to train the kth enterprise local identification model.

[0117] The processing module 403 is configured to update the global parameter according to the M groups of gradient information. If the accuracy of the industry global identification model determined by the updated global parameter is greater than or equal to an accuracy threshold, the industry global identification model is determined according to the updated global parameter.

[0118] Optionally, the receiving module 401 and the sending module 402 can be integrated into a transceiver module. The transceiver module is configured to implement the transceiving function of the Internet of Things card identification apparatus 400.

[0119] In addition, the technical effects of the Internet of Things card identification apparatus 400 can refer to the technical effects of the Internet of Things card identification method shown in Figure 2 or Figure 3 , which will not be repeated here.

[0120] Figure 5 Another structural schematic diagram of an Internet of Things card identification apparatus provided by the embodiment is shown in the figure. The apparatus is applicable to the Internet of Things card identification system shown in Figure 1 . As shown in Figure 5 , the Internet of Things card identification apparatus 500 includes a receiving module 501 and a processing module 502.

[0121] The receiving module 501 is configured to receive an industry global identification model from a service node.

[0122] The industry global identification model is configured to identify an Internet of Things card. The industry global identification model is obtained according to at least one of the following: first sample data or M sets of gradient information. The first sample data is obtained according to M sets of second sample data. The kth set of second sample data in the M sets of second sample data is obtained according to third sample data and the kth set of fourth sample data. The third sample data is data corresponding to M enterprise nodes in user data recorded by an operator node. The kth set of fourth sample data is user data of the kth enterprise node in the M enterprise nodes. The M enterprise nodes belong to the same industry. k and M are positive integers, and 1 < k ≤ M. The kth set of gradient information in the M sets of gradient information is obtained by training the kth enterprise local identification model according to the kth set of second sample data. The kth enterprise local identification model corresponds to the kth enterprise node.

[0123] The processing module 502 is configured to train the industry global identification model according to enterprise individual characteristics to obtain an enterprise individualized identification model.

[0124] Optionally, the Internet of Things card identification apparatus 500 can further include a sending module 503. The sending module is configured to implement the sending function of the Internet of Things card identification apparatus 500.

[0125] Optionally, the receiving module 501 and the sending module 503 can be integrated into a transceiver module. The transceiver module is configured to implement the transceiving function of the Internet of Things card identification apparatus 500.

[0126] In addition, the technical effects of the Internet of Things card identification apparatus 500 can refer to the technical effects of the Internet of Things card identification method shown in Figure 2 or Figure 3 , which will not be repeated here.

[0127] Figure 6 Another Internet of Things card identification device is provided for the embodiment. As shown in the figure, the Internet of Things card identification device 600 includes a processor 601, a memory 602, a communication interface 603 and a bus 604. The processor 601, the memory 602 and the communication interface 603 can be connected through the bus 604. Figure 6

[0128] The processor 601 is the control center of the Internet of Things card identification device, which can be one processor or a general term of multiple processing elements. For example, the processor 601 can be a general central processing unit (CPU), or other general-purpose processors, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0129] As an embodiment, the processor 601 can include one or more CPUs, such as the CPU 0 and the CPU 1 shown in the figure. Figure 6

[0130] The memory 602 can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a magnetic disk storage medium or other magnetic storage devices, or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and can be accessed by a computer, but not limited to this.

[0131] In one possible implementation, the memory 602 can exist independently of the processor 601, and the memory 602 can be connected to the processor 601 through the bus 604, for storing instructions or program codes. When the processor 601 calls and executes the instructions or program codes stored in the memory 602, the Internet of Things card identification method provided by the embodiment can be implemented.

[0132] In another possible implementation, the memory 602 can also be integrated with the processor 601.

[0133] ​​Communication interface 603 is used to connect with other devices via a communication network. The communication network can be Ethernet, a wireless access network, a wireless local area network (WLAN), etc. Communication interface 603 may include a receiving unit for receiving data and a transmitting unit for sending data.

[0134] Bus 604 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0135] It should be pointed out that, Figure 6 The structure shown does not constitute a limitation on the IoT card identification device, except Figure 6 In addition to the components shown, the IoT card identification device 600 may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0136] This application also provides a computer-readable storage medium storing computer instructions, which, when executed on a computer, cause the computer to perform the IoT card identification method as described in the above embodiments.

[0137] The computer readable storage medium, for example, can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a register, a hard disk, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing, or any other medium from which a computer can read instructions. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. Of course, the storage medium can be a part of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). In an embodiment of the present application, the computer readable storage medium can be any tangible medium that can contain or store program codes in the form of instructions or data structures and that can be accessed by the instruction execution system, apparatus, or device.

[0138] The embodiment of the present application further provides a computer program product, which can be directly loaded into the memory and contains software codes, and the computer program product can realize each step of the above-mentioned IoT card identification method and the IoT card identification device after being loaded and executed by a computer.

[0139] In the above embodiments, all or part can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer executes the computer instructions, all or part generates the processes or functions according to the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or data storage device such as one or more servers, data centers, etc. integrated with one or more media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as DVD), or semiconductor media (such as solid state disk (SSD)) and the like.

[0140] Through the description of the above embodiments, those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above division of functional modules is taken as an example, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.

[0141] In several embodiments provided in the present application, it should be understood that the disclosed devices 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 or units is only a logical function division, and actual implementation can have another division manner. For example, a plurality of units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed mutual units can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms. The units shown as separate components can be or can not be physically separated, and the components shown as units can be one physical unit or a plurality of physical units, that is, can be located in one place, or can be distributed in a plurality of different places. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiments.

[0142] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit. When the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions to make a device (which can be a single-chip microcomputer, a chip, etc.) or a processor execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk or an optical disk and various storage medium capable of storing program codes.

[0143] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An Internet of Things card identification method, characterized in that, The method comprises: receiving first sample data from an operator node; wherein the first sample data is obtained according to M groups of second sample data; the kth group of second sample data in the M groups of second sample data is obtained according to the same user data in third sample data and the kth group of fourth sample data; the third sample data is data corresponding to M enterprise nodes in user data recorded by the operator node, and the kth group of fourth sample data is user data of a kth enterprise node in the M enterprise nodes; the M enterprise nodes belong to the same industry; k and M are positive integers, and 1 < k ≤ M; the first sample data is obtained according to industry common characteristics extracted from the M groups of second sample data; receiving M groups of gradient information from the M enterprise nodes; wherein the kth group of gradient information in the M groups of gradient information is obtained by training a kth enterprise local identification model according to the kth group of second sample data; the kth enterprise local identification model corresponds to the kth enterprise node; sending an industry global identification model to the M enterprise nodes; the industry global identification model is used for identifying Internet of Things cards; the industry global identification model is obtained according to at least one of the following: the first sample data or the M groups of gradient information; sending a global parameter to the M enterprise nodes; the global parameter is obtained by training the industry global identification model according to the first sample data; receiving the M groups of gradient information from the M enterprise nodes; the kth group of gradient information in the M groups of gradient information is obtained by training the kth enterprise local identification model according to the global parameter and the kth group of second sample data; updating the global parameter according to the M groups of gradient information; if the accuracy of the industry global identification model determined by the updated global parameter is greater than or equal to an accuracy threshold, determining the industry global identification model according to the updated global parameter.

2. An Internet of Things card identification method, characterized in that, The method comprises: receiving an industry global identification model from a service node; the industry global identification model is used for identifying Internet of Things cards; the industry global identification model is obtained according to at least one of the following: first sample data or M groups of gradient information; the first sample data is obtained according to industry common characteristics extracted from M groups of second sample data; wherein the first sample data is obtained according to M groups of second sample data; the kth group of second sample data in the M groups of second sample data is obtained according to the same user data in third sample data and the kth group of fourth sample data; the third sample data is data corresponding to M enterprise nodes in user data recorded by an operator node, and the kth group of fourth sample data is data of a kth enterprise node in the M enterprise nodes; the M enterprise nodes belong to the same industry; k and M are positive integers, and 1 < k ≤ M; the kth group of gradient information in the M groups of gradient information is obtained by training a kth enterprise local identification model according to the kth group of second sample data; the kth enterprise local identification model corresponds to the kth enterprise node; The industry global recognition model is trained according to the individual characteristics of the enterprises, and an enterprise individualized recognition model is obtained.

3. An Internet of Things card identification device, characterized by, The device comprises a receiving module and a sending module; The receiving module is configured to receive first sample data from an operator node, wherein the first sample data is obtained according to M groups of second sample data, the kth group of second sample data in the M groups of second sample data is obtained according to the same user data in third sample data and the kth group of fourth sample data, the third sample data is data corresponding to M enterprise nodes in user data recorded by the operator node, the kth group of fourth sample data is user data of a kth enterprise node in the M enterprise nodes, the M enterprise nodes belong to the same industry, k and M are positive integers, 1 < k ≤ M, and the first sample data is obtained according to industry common characteristics extracted from the M groups of second sample data; The receiving module is further configured to receive M groups of gradient information from the M enterprise nodes, wherein the kth group of gradient information in the M groups of gradient information is obtained by training a kth enterprise local recognition model according to the kth group of second sample data, and the kth enterprise local recognition model corresponds to the kth enterprise node; The sending module is configured to send an industry global recognition model to the M enterprise nodes, the industry global recognition model is used to identify Internet of Things cards, and the industry global recognition model is obtained according to at least one of the following: the first sample data or the M groups of gradient information; The device further comprises a processing module; The sending module is configured to send a global parameter to the M enterprise nodes, and the global parameter is obtained by training the industry global recognition model according to the first sample data; The receiving module is configured to receive the M groups of gradient information from the M enterprise nodes, and the kth group of gradient information in the M groups of gradient information is obtained by training the kth enterprise local recognition model according to the global parameter and the kth group of second sample data; The processing module is configured to update the global parameter according to the M groups of gradient information; If the accuracy of the industry global recognition model determined by the updated global parameter is greater than or equal to an accuracy threshold, the industry global recognition model is determined according to the updated global parameter.

4. An Internet of Things card identification device, characterized by, The device comprises a receiving module and a processing module; The receiving module is configured to receive an industry global recognition model from a service node, the industry global recognition model is used to identify Internet of Things cards, and the industry global recognition model is obtained according to at least one of the following: first sample data or M groups of gradient information, and the first sample data is obtained according to industry common characteristics extracted from M groups of second sample data; The first sample data is obtained according to M groups of second sample data; kth group of second sample data in the M groups of second sample data is obtained according to third sample data and data of the same user in the kth group of fourth sample data; the third sample data is data corresponding to the M enterprise nodes in the user data recorded by the operator node; the kth group of fourth sample data is user data of the kth enterprise node in the M enterprise nodes; the M enterprise nodes belong to the same industry; k and M are positive integers, 1 < k <= M; kth group of gradient information in the M groups of gradient information is obtained by training the kth enterprise local identification model according to the kth group of second sample data; the kth enterprise local identification model corresponds to the kth enterprise node; The processing module is configured to train the industry global identification model according to the enterprise individual characteristics to obtain an enterprise individualized identification model.

5. An Internet of Things card identification device, characterized by, The memory is configured to store computer execution instructions, and the processor is connected with the memory through a bus; When the device is running, the processor executes the computer execution instructions stored in the memory, so that the device executes the Internet of Things card identification method as claimed in claim 1 or 2.

6. A computer-readable storage medium, characterized in that, The computer instructions are stored, and when the computer instructions run on the computer, the computer executes the Internet of Things card identification method as claimed in claim 1 or 2.

Images