Group Classification Method, Device, Equipment and Storage Medium

Abstract

The present invention relates to artificial intelligence, and provides a group classification method, apparatus, device and storage medium. The method constructs a group augmented matrix including a group matrix according to user feature information and a preset value, performs disassembling and augmentation processing on the group matrix to obtain an initial mapping matrix and an initial feature matrix, constructs a first mask matrix and a second mask matrix according to the missing information of the user feature information, calculates a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, calculates the initial mapping matrix and the group augmented matrix based on the second mask matrix to obtain an intermediate feature matrix, adjusts the intermediate feature matrix based on the relationship matrix between the group matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix, and classifies multiple users based on the target feature matrix, thereby improving the accuracy of the group type. In addition, the present invention also relates to blockchain technology, and the group type can be stored in the blockchain.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a group classification method, device, equipment and storage medium. Background Art

[0002] Currently, when classifying user groups, existing classification algorithms or clustering algorithms are usually used to perform clustering analysis on user information. However, this solution cannot handle scenarios where user information is missing, resulting in inaccurate classification of user groups. Summary of the Invention

[0003] In view of the above, it is necessary to provide a group classification method, device, equipment and storage medium that can solve the technical problem of inaccurate classification of user groups due to missing user information.

[0004] On the one hand, the present invention proposes a group classification method, and the group classification method includes:

[0005] Obtain the user feature information of multiple users in the user group;

[0006] Construct a group augmented matrix according to the user feature information and a preset value, and the group augmented matrix includes a group matrix;

[0007] Perform a disassembling and augmenting process on the group matrix to obtain an initial mapping matrix and an initial feature matrix;

[0008] Construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on a preset feature label;

[0009] Calculate a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculate the initial mapping matrix and the group augmented matrix based on the second mask matrix to obtain an intermediate feature matrix;

[0010] Adjust the intermediate feature matrix based on the relationship matrix between the group matrix and the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix;

[0011] Classify the multiple users based on the target feature matrix to obtain the group type of each user.

[0012] According to a preferred embodiment of the present invention, the constructing a group augmented matrix according to the user feature information and a preset value, and the group augmented matrix includes a group matrix includes:

[0013] Quantify the user characteristic information to obtain the information values of each user on multiple preset characteristic tags;

[0014] Construct the population matrix according to the information values corresponding to each user;

[0015] Generate the population augmented matrix according to the population matrix and the row vector corresponding to the preset value.

[0016] According to a preferred embodiment of the present invention, the disassembling and augmenting process of the population matrix to obtain the initial mapping matrix and the initial feature matrix includes:

[0017] Disassemble the population matrix to obtain a first disassembled matrix and a second disassembled matrix;

[0018] Perform an augmenting process on the first disassembled matrix to obtain the initial mapping matrix;

[0019] Perform an augmenting process on the transposed matrix of the second disassembled matrix to obtain the initial feature matrix.

[0020] According to a preferred embodiment of the present invention, the construction of the first mask matrix of the initial mapping matrix and the second mask matrix of the initial feature matrix according to the missing information of the user characteristic information on the preset characteristic tags includes:

[0021] Identify the first representation position of the missing information in the first disassembled matrix, and determine the matrix positions in the first disassembled matrix except the first representation position as the second representation positions;

[0022] Based on the first configuration value, replace the elements corresponding to the first representation position in the first disassembled matrix, and based on the second configuration value, replace the elements corresponding to the second representation position in the first disassembled matrix to obtain the first mask matrix;

[0023] Identify the third representation position of the missing information in the second disassembled matrix, and determine the matrix positions in the second disassembled matrix except the third representation position as the fourth representation positions;

[0024] Based on the first configuration value, replace the elements corresponding to the third representation position in the second disassembled matrix, and based on the second configuration value, replace the elements corresponding to the fourth representation position in the second disassembled matrix to obtain the second mask matrix.

[0025] According to a preferred embodiment of the present invention, the calculation of the target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain the intermediate mapping matrix includes:

[0026] Augment the transposed matrix of the group matrix to obtain the target matrix;

[0027] Identify the target number of rows and the target number of columns in the first mask matrix corresponding to the second configuration value;

[0028] Extract feature columns from the initial feature matrix based on the target number of rows and the target number of columns, and extract target columns from the target matrix based on the target number of rows and the target number of columns;

[0029] Perform non - negative least squares operation on the feature columns and the target columns to obtain the intermediate mapping matrix.

[0030] According to a preferred embodiment of the present invention, adjusting the intermediate feature matrix based on the relationship matrix between the group matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix includes:

[0031] Calculate the product of the intermediate mapping matrix and the intermediate feature matrix to obtain the relationship matrix;

[0032] Calculate the root - mean - square error between the relationship matrix and the group matrix;

[0033] Adjust the intermediate feature matrix based on the root - mean - square error until the root - mean - square error is minimized to obtain the target feature matrix.

[0034] According to a preferred embodiment of the present invention, classifying the multiple users based on the target feature matrix to obtain the group type of each user includes:

[0035] If the number of rows of the target feature matrix is greater than a preset row - number threshold, obtain multiple user types corresponding to each user from the target feature matrix;

[0036] For each user, if there are multiple types among the multiple user types, count the number of user types corresponding to the same type;

[0037] Determine the user type corresponding to the largest number of types as the group type of the user.

[0038] On the other hand, the present invention also proposes a group classification device, and the group classification device includes:

[0039] An acquisition unit for acquiring user feature information of multiple users in a user group;

[0040] A construction unit for constructing a group augmented matrix according to the user feature information and a preset value, and the group augmented matrix includes a group matrix;

[0041] An augmentation unit, configured to perform disassembling and augmenting processing on the population matrix to obtain an initial mapping matrix and an initial feature matrix;

[0042] The construction unit is further configured to construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on the preset feature tags;

[0043] A calculation unit, configured to calculate a target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculate the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix;

[0044] An adjustment unit, configured to adjust the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix;

[0045] A classification unit, configured to classify the multiple users based on the target feature matrix to obtain the population type of each user.

[0046] On the other hand, the present invention further provides an electronic device, where the electronic device includes:

[0047] A memory, storing computer-readable instructions; and

[0048] A processor, configured to execute the computer-readable instructions stored in the memory to implement the population classification method.

[0049] On the other hand, the present invention further provides a computer-readable storage medium, where computer-readable instructions are stored in the computer-readable storage medium, and the computer-readable instructions are executed by a processor in an electronic device to implement the population classification method.

[0050] It can be seen from the above technical solutions that the present application can avoid the influence brought by the missing information of the user feature information by constructing the first mask matrix and the second mask matrix, and can improve the generation accuracy of the target feature matrix through the decomposition, update and adjustment of the non-missing values of the user feature information on the intermediate mapping matrix and the intermediate feature matrix, thereby improving the classification accuracy of the user population. Description of the Drawings

[0051] Figure 1 is a flowchart of a preferred embodiment of the population classification method of the present invention.

[0052] Figure 2 is a schematic structural diagram of a target feature matrix in the present invention.

[0053] Figure 3It is another structural schematic diagram of the target feature matrix in the present invention.

[0054] Figure 4 It is a functional module diagram of a preferred embodiment of the group classification device of the present invention.

[0055] Figure 5 It is a structural schematic diagram of an electronic device of a preferred embodiment for implementing the group classification method of the present invention. Detailed implementation manners

[0056] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] As Figure 1 shown, it is a flowchart of a preferred embodiment of the group classification method of the present invention. According to different requirements, the order of steps in this flowchart can be changed and some steps can be omitted.

[0058] The group classification method can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results.

[0059] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0060] The group classification method is applied to one or more electronic devices. The electronic device is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored computer-readable instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0061] The electronic device can be any kind of electronic product that can perform human-computer interaction with users. For example, a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an Internet Protocol Television (IPTV), a smart wearable device, etc.

[0062] The electronic device may include a network device and / or a user device. Among them, the network device includes, but is not limited to, a single network electronic device, a group of electronic devices composed of multiple network electronic devices, or a cloud composed of a large number of hosts or network electronic devices based on cloud computing.

[0063] The network where the electronic device is located includes, but is not limited to: the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), etc.

[0064] 101. Obtain the user characteristic information of multiple users in the user group.

[0065] In at least one embodiment of the present invention, the user group may be a group composed of all permanent residents or household registered users in a certain area, etc. The multiple users refer to the internal users in the user group, and the multiple users may include users in multiple age stages.

[0066] The user characteristic information refers to the information corresponding to each user on a preset characteristic label. Among them, the preset characteristic label may include, but is not limited to: labels such as age, occupation, and interest.

[0067] In at least one embodiment of the present invention, the electronic device obtains the information corresponding to each user on multiple preset characteristic labels from the database corresponding to the user group as the user characteristic information.

[0068] 102. Construct a group augmented matrix according to the user characteristic information and a preset value. The group augmented matrix includes a group matrix.

[0069] In at least one embodiment of the present invention, the preset value can be set according to actual needs. In order to avoid the influence of the preset value on the classification of the multiple users, the multiple preset values in the group augmented matrix are equal. For example, the multiple preset values are all 0.

[0070] The population matrix refers to the representation of the user feature information corresponding to the user population. For example, if the number of users in the multiple users is n and the number of tags of the preset feature tags is m, the representation form of the population matrix is m*n.

[0071] In at least one embodiment of the present invention, the electronic device constructs a population augmented matrix according to the user feature information and a preset value. The population augmented matrix includes a population matrix including:

[0072] Perform quantization processing on the user feature information to obtain information values of each user on multiple preset feature tags;

[0073] Construct the population matrix according to the information values corresponding to each user;

[0074] Generate the population augmented matrix according to the population matrix and the row vector corresponding to the preset value.

[0075] For example, the representation form of the population matrix is m*n, and the row vector corresponding to the preset value is 1*n, then the representation form of the population augmented matrix is (m + 1)*n.

[0076] By constructing a preset value with equal values as the row vector, the influence of preset values with unequal values on the classification of the multiple users can be avoided. Furthermore, by augmenting the population matrix through the row vector, the situation that the multiple users cannot be classified due to missing information in the user feature information can be avoided.

[0077] 103. Perform disassembling and augmenting processing on the population matrix to obtain an initial mapping matrix and an initial feature matrix.

[0078] In at least one embodiment of the present invention, the initial mapping matrix and the initial feature matrix respectively refer to the matrices obtained after performing disassembling and augmenting processing on the population matrix.

[0079] In at least one embodiment of the present invention, the electronic device performs disassembling and augmenting processing on the population matrix to obtain an initial mapping matrix and an initial feature matrix, including:

[0080] Disassemble the population matrix to obtain a first disassembled matrix and a second disassembled matrix;

[0081] Perform augmenting processing on the first disassembled matrix to obtain the initial mapping matrix;

[0082] Perform augmenting processing on the transposed matrix of the second disassembled matrix to obtain the initial feature matrix.

[0083] For example, the representation form of the first decomposition matrix is m*k, the representation form of the second decomposition matrix is k*m. After augmentation processing, the representation form of the initial mapping matrix is (m + 1)*k, and the representation form of the initial feature matrix is (n + k)*k.

[0084] 104. Construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on the preset feature tags.

[0085] In at least one embodiment of the present invention, the missing information refers to the information that the electronic device cannot obtain for any user on the preset feature tags.

[0086] In at least one embodiment of the present invention, the electronic device constructs the first mask matrix of the initial mapping matrix and the second mask matrix of the initial feature matrix according to the missing information of the user feature information on the preset feature tags, including:

[0087] Identify the first representation position of the missing information in the first decomposition matrix, and determine the matrix positions other than the first representation position in the first decomposition matrix as the second representation positions;

[0088] Based on the first configuration value, replace the elements corresponding to the first representation position in the first decomposition matrix, and based on the second configuration value, replace the elements corresponding to the second representation position in the first decomposition matrix to obtain the first mask matrix;

[0089] Identify the third representation position of the missing information in the second decomposition matrix, and determine the matrix positions other than the third representation position in the second decomposition matrix as the fourth representation positions;

[0090] Based on the first configuration value, replace the elements corresponding to the third representation position in the second decomposition matrix, and based on the second configuration value, replace the elements corresponding to the fourth representation position in the second decomposition matrix to obtain the second mask matrix.

[0091] Wherein, the first configuration value can be set to 0, and the second configuration value can be set to 1.

[0092] By constructing the first mask matrix and the second mask matrix, the electronic device can be assisted to quickly identify non-missing values, and the generation efficiency of the intermediate mapping matrix and the intermediate feature matrix can be improved.

[0093] 105. Calculate the target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculate the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix.

[0094] In at least one embodiment of the present invention, the target matrix refers to the matrix obtained by augmenting the transposed matrix of the population matrix. For example, if the population matrix is in the form of m*n, then the target matrix is in the form of (n + k)*m.

[0095] The intermediate mapping matrix and the intermediate feature matrix respectively refer to the matrices disassembled when there are missing values in the population matrix.

[0096] In at least one embodiment of the present invention, when the electronic device calculates the target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, it includes:

[0097] Augment the transposed matrix of the population matrix to obtain the target matrix;

[0098] Identify the target row number and target column number corresponding to the second configuration value in the first mask matrix;

[0099] Based on the target row number and the target column number, extract the feature columns from the initial feature matrix, and based on the target row number and the target column number, extract the target columns from the target matrix;

[0100] Perform non - negative least squares operation on the feature columns and the target columns to obtain the intermediate mapping matrix.

[0101] Through the second configuration value, the feature columns and the target columns can be quickly extracted, thereby improving the generation efficiency of the intermediate mapping matrix.

[0102] In at least one embodiment of the present invention, when the electronic device calculates the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix, it includes:

[0103] Identify the initial row number and initial column number corresponding to the second configuration value in the second mask matrix;

[0104] Based on the initial row number and the initial column number, extract the mapping rows from the initial mapping matrix, and based on the initial row number and the initial column number, extract the population rows from the population augmentation matrix;

[0105] Perform non - negative least - squares operation on the mapping row and the population row to obtain the intermediate feature matrix.

[0106] 106. Adjust the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix.

[0107] In at least one embodiment of the present invention, the relationship matrix refers to the product of the intermediate mapping matrix and the intermediate feature matrix.

[0108] The target feature matrix refers to the adjusted intermediate feature matrix when the root - mean - square error between the relationship matrix and the population matrix is the smallest. The form of the target feature matrix is k*n, where n represents the number of users among the multiple users, and k represents the number of types of population types. As Figure 2 shown, Figure 2 is a schematic structural diagram of the target feature matrix in the present invention. Among them, the form of the target feature matrix is 3*n. The target feature matrix includes the population types of n users and 3 types of population types. Specifically, Figure 2 the population type corresponding to user 1 is type B, the population type corresponding to user 2 is type A, the population type corresponding to user 3 is type C, and the population type corresponding to user n is type C.

[0109] In at least one embodiment of the present invention, the electronic device adjusts the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix, including:

[0110] Calculate the product of the intermediate mapping matrix and the intermediate feature matrix to obtain the relationship matrix;

[0111] Calculate the root - mean - square error between the relationship matrix and the population matrix;

[0112] Adjust the intermediate feature matrix based on the root - mean - square error until the root - mean - square error is the smallest to obtain the target feature matrix.

[0113] Through the above - mentioned implementation manner, the intermediate feature matrix can be iteratively adjusted until the root - mean - square error is the smallest, improving the accuracy of the target feature matrix.

[0114] 107. Classify the multiple users based on the target feature matrix to obtain the population type of each user.

[0115] It should be emphasized that to further ensure the privacy and security of the above - mentioned population type, the above - mentioned population type can also be stored in a node of a blockchain.

[0116] In at least one embodiment of the present invention, the group types may include, but are not limited to: sports, shopping, etc.

[0117] In at least one embodiment of the present invention, the electronic device classifies the multiple users based on the target feature matrix to obtain the group type of each user, including:

[0118] If the number of rows of the target feature matrix is greater than the preset row number threshold, obtain the multiple user types corresponding to each user from the target feature matrix;

[0119] For each user, if there are multiple types of the multiple user types, count the number of types of the user types corresponding to the same type;

[0120] Determine the user type corresponding to the largest number of types as the group type of the user.

[0121] Among them, the preset row number threshold is usually set to 1.

[0122] As Figure 3 shown, Figure 3 is another structural schematic diagram of the target feature matrix in the present invention. Figure 3 For user 1, the corresponding user types are two types, group type A and group type C. Since the number of types of user 1 in group type A is the largest, the group type corresponding to user 1 is group type A. According to the above embodiment, other users are type-identified, and the group type corresponding to user 2 is group type C, the group type corresponding to user 3 is group type B, and the group type corresponding to user n is group type B.

[0123] Through the above implementation manner, when the number of rows of the target feature matrix is greater than the preset row number threshold and there are multiple types of multiple user types for any user, the group type of each user can be accurately determined.

[0124] In other embodiments, if the number of rows of the target feature matrix is equal to the preset row number threshold, the type corresponding to each matrix element in the target feature matrix is determined as the group type of each user.

[0125] Through the above implementation manner, the group type can be directly determined.

[0126] As can be seen from the above technical solutions, by constructing the first mask matrix and the second mask matrix, the present application can avoid the influence brought by the missing information in the user feature information. By decomposing, updating and adjusting the intermediate mapping matrix and the intermediate feature matrix with the non-missing values of the user feature information, the generation accuracy of the target feature matrix can be improved, thereby improving the classification accuracy of the user group.

[0127] As Figure 4 shown, it is a functional module diagram of a preferred embodiment of the group classification device of the present invention. The group classification device 11 includes an acquisition unit 110, a construction unit 111, an augmentation unit 112, a calculation unit 113, an adjustment unit 114, and a classification unit 115. The module / unit referred to in the present invention means a series of computer-readable instruction segments that can be acquired by a processor 13 and can complete fixed functions, and are stored in a memory 12. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0128] The acquisition unit 110 acquires the user feature information of multiple users in the user group.

[0129] In at least one embodiment of the present invention, the user group may be a group composed of all permanent residents or household registered users in a certain area, etc. The multiple users refer to the internal users in the user group, and the multiple users may include users in multiple age groups.

[0130] The user feature information refers to the information corresponding to each user on a preset feature label, where the preset feature label may include, but is not limited to: labels such as age, occupation, and interest.

[0131] In at least one embodiment of the present invention, the acquisition unit 110 acquires the information corresponding to each user on multiple preset feature labels from a database corresponding to the user group as the user feature information.

[0132] The construction unit 111 constructs a group augmentation matrix according to the user feature information and a preset value, and the group augmentation matrix includes a group matrix.

[0133] In at least one embodiment of the present invention, the preset value can be set according to actual needs. In order to avoid the influence of the preset value on the classification of the multiple users, the multiple preset values in the group augmentation matrix are equal. For example, the multiple preset values are all 0.

[0134] The population matrix refers to the representation of the user feature information corresponding to the user population. For example, if the number of users in the multiple users is n and the number of labels of the preset feature labels is m, the representation form of the population matrix is m*n.

[0135] In at least one embodiment of the present invention, the construction unit 111 constructs a population augmented matrix according to the user feature information and a preset value. The population augmented matrix includes a population matrix, including:

[0136] Quantize the user feature information to obtain the information values of each user on multiple preset feature labels;

[0137] Construct the population matrix according to the information values corresponding to each user;

[0138] Generate the population augmented matrix according to the population matrix and the row vector corresponding to the preset value.

[0139] For example, the representation form of the population matrix is m*n, and the row vector corresponding to the preset value is 1*n, then the representation form of the population augmented matrix is (m + 1)*n.

[0140] By constructing a preset value with equal values as the row vector, the influence of the preset values with unequal values on the classification of the multiple users can be avoided. Furthermore, by augmenting the population matrix through the row vector, the situation that the multiple users cannot be classified due to the existence of missing information in the user feature information can be avoided.

[0141] The augmentation unit 112 performs a disassembling and augmenting process on the population matrix to obtain an initial mapping matrix and an initial feature matrix.

[0142] In at least one embodiment of the present invention, the initial mapping matrix and the initial feature matrix respectively refer to the matrices obtained after performing a disassembling and augmenting process on the population matrix.

[0143] In at least one embodiment of the present invention, the augmentation unit 112 performs a disassembling and augmenting process on the population matrix to obtain an initial mapping matrix and an initial feature matrix, including:

[0144] Disassemble the population matrix to obtain a first disassembled matrix and a second disassembled matrix;

[0145] Perform an augmenting process on the first disassembled matrix to obtain the initial mapping matrix;

[0146] Perform an augmenting process on the transposed matrix of the second disassembled matrix to obtain the initial feature matrix.

[0147] For example, the representation form of the first disassembly matrix is m*k, and the representation form of the second disassembly matrix is k*m. After augmentation processing, the representation form of the initial mapping matrix is (m + 1)*k, and the representation form of the initial feature matrix is (n + k)*k.

[0148] The construction unit 111 constructs a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on the preset feature label.

[0149] In at least one embodiment of the present invention, the missing information refers to the information that the electronic device cannot obtain for any user on the preset feature label.

[0150] In at least one embodiment of the present invention, the construction unit 111 constructing the first mask matrix of the initial mapping matrix and the second mask matrix of the initial feature matrix according to the missing information of the user feature information on the preset feature label includes:

[0151] Identify the first representation position of the missing information in the first disassembly matrix, and determine the matrix positions other than the first representation position in the first disassembly matrix as the second representation position;

[0152] Based on the first configuration value, replace the elements corresponding to the first representation position in the first disassembly matrix, and based on the second configuration value, replace the elements corresponding to the second representation position in the first disassembly matrix to obtain the first mask matrix;

[0153] Identify the third representation position of the missing information in the second disassembly matrix, and determine the matrix positions other than the third representation position in the second disassembly matrix as the fourth representation position;

[0154] Based on the first configuration value, replace the elements corresponding to the third representation position in the second disassembly matrix, and based on the second configuration value, replace the elements corresponding to the fourth representation position in the second disassembly matrix to obtain the second mask matrix.

[0155] Wherein, the first configuration value can be set to 0, and the second configuration value can be set to 1.

[0156] By constructing the first mask matrix and the second mask matrix, the electronic device can be assisted in quickly identifying non-missing values, and the generation efficiency of the intermediate mapping matrix and the intermediate feature matrix can be improved.

[0157] The computing unit 113 calculates the target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculates the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix.

[0158] In at least one embodiment of the present invention, the target matrix refers to the matrix obtained by augmenting the transposed matrix of the population matrix. For example, if the population matrix is in the form of m*n, then the target matrix is in the form of (n + k)*m.

[0159] The intermediate mapping matrix and the intermediate feature matrix respectively refer to the matrices disassembled when there are missing values in the population matrix.

[0160] In at least one embodiment of the present invention, the computing unit 113 calculates the target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, including:

[0161] Augment the transposed matrix of the population matrix to obtain the target matrix;

[0162] Identify the target row number and target column number corresponding to the second configuration value in the first mask matrix;

[0163] Based on the target row number and the target column number, extract the feature columns from the initial feature matrix, and based on the target row number and the target column number, extract the target columns from the target matrix;

[0164] Perform non - negative least squares operation on the feature columns and the target columns to obtain the intermediate mapping matrix.

[0165] The second configuration value can quickly extract the feature columns and the target columns, thereby improving the generation efficiency of the intermediate mapping matrix.

[0166] In at least one embodiment of the present invention, the computing unit 113 calculates the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix, including:

[0167] Identify the initial row number and initial column number corresponding to the second configuration value in the second mask matrix;

[0168] Based on the initial row number and the initial column number, extract the mapping rows from the initial mapping matrix, and based on the initial row number and the initial column number, extract the population rows from the population augmentation matrix;

[0169] Perform non - negative least - squares operation on the mapping row and the population row to obtain the intermediate feature matrix.

[0170] The adjustment unit 114 adjusts the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix.

[0171] In at least one embodiment of the present invention, the relationship matrix refers to the product of the intermediate mapping matrix and the intermediate feature matrix.

[0172] The target feature matrix refers to the adjusted intermediate feature matrix when the root - mean - square error between the relationship matrix and the population matrix is the smallest. The form of the target feature matrix is k*n, where n represents the number of users among the multiple users, and k represents the number of types of population types. As Figure 2 shown, Figure 2 is a schematic structural diagram of the target feature matrix in the present invention. Among them, the form of the target feature matrix is 3*n. The target feature matrix includes the population types of n users and 3 types of population types. Specifically, Figure 2 the population type corresponding to user 1 is type B, the population type corresponding to user 2 is type A, the population type corresponding to user 3 is type C, and the population type corresponding to user n is type C.

[0173] In at least one embodiment of the present invention, the adjustment unit 114 adjusts the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix, including:

[0174] Calculate the product of the intermediate mapping matrix and the intermediate feature matrix to obtain the relationship matrix;

[0175] Calculate the root - mean - square error between the relationship matrix and the population matrix;

[0176] Adjust the intermediate feature matrix based on the root - mean - square error until the root - mean - square error is the smallest to obtain the target feature matrix.

[0177] Through the above - mentioned implementation manner, the intermediate feature matrix can be iteratively adjusted until the root - mean - square error is the smallest, improving the accuracy of the target feature matrix.

[0178] The classification unit 115 classifies the multiple users based on the target feature matrix to obtain the population type of each user.

[0179] It should be emphasized that to further ensure the privacy and security of the above - mentioned population types, the above - mentioned population types can also be stored in a node of a blockchain.

[0180] In at least one embodiment of the present invention, the group types may include, but are not limited to: sports, shopping, etc.

[0181] In at least one embodiment of the present invention, the classification unit 115 classifies the multiple users based on the target feature matrix, and the group type obtained for each user includes:

[0182] If the number of rows of the target feature matrix is greater than a preset row number threshold, obtain the multiple user types corresponding to each user from the target feature matrix;

[0183] For each user, if there are multiple types among the multiple user types, count the number of user types corresponding to the same type;

[0184] Determine the user type corresponding to the type with the largest value as the group type of the user.

[0185] Among them, the preset row number threshold is usually set to 1.

[0186] As Figure 3 shown, Figure 3 is another structural schematic diagram of the target feature matrix in the present invention. Figure 3 For user 1, the corresponding user types are group type A and group type C. Since the number of user 1 in group type A is the largest, the group type corresponding to user 1 is group type A. According to the above embodiment, type recognition is performed on other users, and the group type corresponding to user 2 is group type C, the group type corresponding to user 3 is group type B, and the group type corresponding to user n is group type B.

[0187] Through the above implementation manner, when the number of rows of the target feature matrix is greater than the preset row number threshold and there are multiple types among the multiple user types of any user, the group type of each user can be accurately determined.

[0188] In other embodiments, if the number of rows of the target feature matrix is equal to the preset row number threshold, the type corresponding to each matrix element in the target feature matrix is determined as the group type of each user.

[0189] Through the above implementation manner, the group type can be directly determined.

[0190] As can be seen from the above technical solutions, by constructing the first mask matrix and the second mask matrix, the present application can avoid the influence brought by the missing information in the user feature information. By decomposing, updating, and adjusting the intermediate mapping matrix and the intermediate feature matrix with the non-missing values of the user feature information, the generation accuracy of the target feature matrix can be improved, thereby improving the classification accuracy of the user group.

[0191] As Figure 5 shown, it is a schematic structural diagram of an electronic device according to a preferred embodiment of the method for implementing group classification of the present invention.

[0192] In an embodiment of the present invention, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer-readable instructions stored in the memory 12 and executable on the processor 13, such as a group classification program.

[0193] Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 1, and does not constitute a limitation on the electronic device 1. It may include more or fewer components than shown, or combine some components, or different components. For example, the electronic device 1 may further include input / output devices, network access devices, buses, etc.

[0194] The processor 13 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor 13 is the operation core and control center of the electronic device 1, connecting various parts of the entire electronic device 1 through various interfaces and lines, and executing the operating system of the electronic device 1 and various installed application programs, program codes, etc.

[0195] Exemplarily, the computer-readable instructions may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to implement the present invention. The one or more modules / units may be a series of computer-readable instruction segments capable of performing specific functions, and these computer-readable instruction segments are used to describe the execution process of the computer-readable instructions in the electronic device 1. For example, the computer-readable instructions may be divided into an acquisition unit 110, a construction unit 111, an augmentation unit 112, a calculation unit 113, an adjustment unit 114, and a classification unit 115.

[0196] The memory 12 may be used to store the computer-readable instructions and / or modules. By running or executing the computer-readable instructions and / or modules stored in the memory 12, and by invoking the data stored in the memory 12, the processor 13 realizes various functions of the electronic device 1. The memory 12 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area may store data created according to the use of the electronic device. The memory 12 may include non-volatile and volatile memories, such as: hard disks, memories, plug-in hard disks, smart media cards (SMCs), secure digital (SD) cards, flash cards, at least one magnetic disk storage device, flash memory devices, or other storage devices.

[0197] The memory 12 may be an external memory and / or an internal memory of the electronic device 1. Further, the memory 12 may be a memory in physical form, such as a memory stick, a TF card (Trans-flash Card), and so on.

[0198] If the modules / units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above method embodiments of the present invention, it may also be completed by computer-readable instructions instructing related hardware. The computer-readable instructions may be stored in a computer-readable storage medium. When the computer-readable instructions are executed by the processor, the steps of the above method embodiments may be realized.

[0199] Among them, the computer-readable instructions include computer-readable instruction codes, which may be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer-readable instruction codes, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memories, read-only memories (ROMs), and random access memories (RAMs).

[0200] The blockchain referred to in the present invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithms. Blockchain, in essence, is a decentralized database, a series of data blocks generated by using cryptographic methods. Each data block contains information on a batch of network transactions, which is used to verify the validity (anti-counterfeiting) of the information and generate the next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, etc.

[0201] Combined Figure 1 , the memory 12 in the electronic device 1 stores computer-readable instructions to implement a group classification method, and the processor 13 can execute the computer-readable instructions to implement:

[0202] Obtain the user feature information of multiple users in the user group;

[0203] Construct a group augmented matrix according to the user feature information and a preset value. The group augmented matrix includes a group matrix;

[0204] Perform disassembling and augmenting processing on the group matrix to obtain an initial mapping matrix and an initial feature matrix;

[0205] Construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on preset feature tags;

[0206] Calculate a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculate the initial mapping matrix and the group augmented matrix based on the second mask matrix to obtain an intermediate feature matrix;

[0207] Adjust the intermediate feature matrix based on the relationship matrix between the group matrix, the intermediate mapping matrix, and the intermediate feature matrix to obtain a target feature matrix;

[0208] Classify the multiple users based on the target feature matrix to obtain the group type of each user.

[0209] Specifically, for the specific implementation method of the above computer-readable instructions by the processor 13, reference may be made to Figure 1 the description of the relevant steps in the corresponding embodiments, which will not be elaborated here.

[0210] In several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0211] Computer-readable instructions are stored on the computer-readable storage medium, wherein when the computer-readable instructions are executed by the processor 13, the following steps are implemented:

[0212] Obtain user characteristic information of multiple users in the user group;

[0213] Construct a group augmented matrix according to the user characteristic information and a preset value, where the group augmented matrix includes a group matrix;

[0214] Perform a disassembling and augmenting process on the group matrix to obtain an initial mapping matrix and an initial feature matrix;

[0215] Construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user characteristic information on the preset feature tags;

[0216] Calculate a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculate the initial mapping matrix and the group augmented matrix based on the second mask matrix to obtain an intermediate feature matrix;

[0217] Adjust the intermediate feature matrix based on the relationship matrix between the group matrix and the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix;

[0218] Classify the multiple users based on the target feature matrix to obtain the group type of each user.

[0219] 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 may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0220] In addition, in each embodiment of the present invention, each functional module may be integrated into one processing unit, may exist separately as individual units physically, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0221] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Thus, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed within the present invention. Any reference signs in the claims should not be construed as limiting the claims concerned.

[0222] In addition, it is obvious that the term "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The described multiple units or devices may also be implemented by one unit or device through software or hardware. The terms such as first and second are used to denote names and do not represent any particular order.

Claims

1. A method for group classification, characterized in that, The group classification method includes: Obtaining user feature information of multiple users in a user group; Constructing a group augmented matrix according to the user feature information and a preset value, where the group augmented matrix includes a group matrix; Performing a disassembling and augmenting process on the group matrix to obtain an initial mapping matrix and an initial feature matrix; Constructing a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on preset feature labels, including: identifying a first representation position of the missing information in a first disassembled matrix, and determining matrix positions in the first disassembled matrix other than the first representation position as second representation positions; based on a first configuration value, replacing elements corresponding to the first representation position in the first disassembled matrix, and based on a second configuration value, replacing elements corresponding to the second representation positions in the first disassembled matrix to obtain the first mask matrix; identifying a third representation position of the missing information in a second disassembled matrix, and determining matrix positions in the second disassembled matrix other than the third representation position as fourth representation positions; based on the first configuration value, replacing elements corresponding to the third representation position in the second disassembled matrix, and based on the second configuration value, replacing elements corresponding to the fourth representation positions in the second disassembled matrix to obtain the second mask matrix; Calculating a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculating the initial mapping matrix and the group augmented matrix based on the second mask matrix to obtain an intermediate feature matrix; Adjusting the intermediate feature matrix based on a relationship matrix between the group matrix and the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix; Classifying the multiple users based on the target feature matrix to obtain the group type of each user.

2. The group classification method according to claim 1, wherein The constructing a group augmented matrix according to the user feature information and a preset value, where the group augmented matrix includes a group matrix includes: Performing quantization processing on the user feature information to obtain information values of each user on multiple preset feature labels; Constructing the group matrix according to the information values corresponding to each user; Generating the group augmented matrix according to the group matrix and a row vector corresponding to the preset value.

3. The group classification method according to claim 1, wherein The performing a disassembling and augmenting process on the group matrix to obtain an initial mapping matrix and an initial feature matrix includes: Disassembling the group matrix to obtain the first disassembled matrix and the second disassembled matrix; Performing an augmenting process on the first disassembled matrix to obtain the initial mapping matrix; Performing an augmenting process on a transposed matrix of the second disassembled matrix to obtain the initial feature matrix.

4. The group classification method according to claim 1, wherein The calculating a target matrix corresponding to the group matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix includes: Performing an augmenting process on a transposed matrix of the group matrix to obtain the target matrix; Identify the target number of rows and the target number of columns in the first mask matrix corresponding to the second configuration value; Based on the target number of rows and the target number of columns, extract feature columns from the initial feature matrix, and based on the target number of rows and the target number of columns, extract target columns from the target matrix; Perform non - negative least squares operation on the feature columns and the target columns to obtain the intermediate mapping matrix.

5. The group classification method according to claim 1, characterized in that, The adjusting the intermediate feature matrix based on the relationship matrix between the population matrix, the intermediate mapping matrix and the intermediate feature matrix to obtain the target feature matrix includes: Calculate the product of the intermediate mapping matrix and the intermediate feature matrix to obtain the relationship matrix; Calculate the root mean square error between the relationship matrix and the population matrix; Adjust the intermediate feature matrix based on the root mean square error until the root mean square error is minimized to obtain the target feature matrix.

6. The group classification method according to claim 1, wherein The classifying the multiple users based on the target feature matrix to obtain the population type of each user includes: If the number of rows of the target feature matrix is greater than a preset row - number threshold, obtain multiple user types corresponding to each user from the target feature matrix; For each user, if there are multiple types among the multiple user types, count the number of user types corresponding to the same type; Determine the user type corresponding to the largest number of types as the population type of the user.

7. A group classification device, characterized in that, The population classification device includes: An acquisition unit, configured to acquire user feature information of multiple users in a user population; A construction unit, configured to construct a population augmented matrix according to the user feature information and a preset value, where the population augmented matrix includes a population matrix; An augmentation unit, configured to perform disassembling and augmentation processing on the population matrix to obtain an initial mapping matrix and an initial feature matrix; The construction unit is further configured to construct a first mask matrix of the initial mapping matrix and a second mask matrix of the initial feature matrix according to the missing information of the user feature information on a preset feature label, including: identifying a first representation position of the missing information in a first disassembled matrix, and determining matrix positions in the first disassembled matrix except the first representation position as second representation positions; based on a first configuration value, replacing elements in the first disassembled matrix corresponding to the first representation position, and based on a second configuration value, replacing elements in the first disassembled matrix corresponding to the second representation positions to obtain the first mask matrix; identifying a third representation position of the missing information in a second disassembled matrix, and determining matrix positions in the second disassembled matrix except the third representation position as fourth representation positions; based on the first configuration value, replacing elements in the second disassembled matrix corresponding to the third representation position, and based on the second configuration value, replacing elements in the second disassembled matrix corresponding to the fourth representation positions to obtain the second mask matrix; A calculation unit for calculating a target matrix corresponding to the population matrix and the initial feature matrix based on the first mask matrix to obtain an intermediate mapping matrix, and calculating the initial mapping matrix and the population augmentation matrix based on the second mask matrix to obtain an intermediate feature matrix; An adjustment unit for adjusting the intermediate feature matrix based on the relationship matrix between the population matrix and the intermediate mapping matrix and the intermediate feature matrix to obtain a target feature matrix; A classification unit for classifying the multiple users based on the target feature matrix to obtain the population type of each user.

8. An electronic device, characterized in that, The electronic device includes: A memory storing computer-readable instructions; and A processor for executing the computer-readable instructions stored in the memory to implement the population classification method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored in the computer-readable storage medium, and the computer-readable instructions are executed by a processor in the electronic device to implement the population classification method according to any one of claims 1 to 6.

Images