User scoring method and device, electronic equipment and computer storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a user scoring method, which comprises the steps of obtaining user feature dimensions of a target activity, classifying the user feature dimensions, and obtaining user feature data corresponding to each user feature dimension; inputting each user feature dimension and the user feature data thereof into a preset analytic hierarchy model for processing to obtain a dimension judgment matrix and a feature judgment matrix; processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score; obtaining the information entropy weight of each candidate user, and evaluating the user data according to the information entropy weight to obtain a second user score; and according to the first user score and the second user score of each candidate user, determining a target user and pushing a target activity to the target user. The invention further discloses a user scoring device, electronic equipment and a computer storage medium. According to the invention, the accuracy of user evaluation in an unsupervised scene is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a user rating method, device, electronic device and computer storage medium. Background Art

[0002] Currently, with the continuous emergence of new things, it is often difficult to accurately screen and recommend them to users in the absence of prior knowledge.

[0003] In unsupervised scenarios, the current method of making recommendations mainly involves randomly selecting users, which requires a long period of user experiments to obtain recommendation feedback and adjust the recommendations. This user screening method has a long experimental cycle and slow iteration. In addition to the method of screening users through random selection, there is also a method of formulating relevant rules based on historical experience, screening users according to the relevant rules, and then making recommendations. However, the various rules of this recommendation method are equal, making it difficult to merge the rules, and once the rules are determined, the scope of the customer group will also be determined. As a result, it ultimately leads to the technical problem of being unable to accurately evaluate and screen users in unsupervised scenarios. Summary of the invention

[0004] The embodiments of the present application provide a user rating method, device, electronic device and computer storage medium, which can solve the current technical problem of being unable to accurately evaluate and screen users.

[0005] The present application embodiment provides a user rating method, including:

[0006] Acquire user feature dimensions of the target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions;

[0007] Input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively;

[0008] Processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user;

[0009] Obtaining an information entropy weight of each candidate user, and evaluating user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user;

[0010] A target user is determined according to the first user score and the second user score of each candidate user, and the target activity is pushed to the target user.

[0011] Furthermore, each of the user feature dimensions and the user feature data are input into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, including:

[0012] Based on the hierarchical analysis model, the user feature dimension and the user feature data are processed respectively to obtain first feature data corresponding to the user feature dimension and second feature data corresponding to the user feature data;

[0013] Obtain a mapping relationship table of preset feature data and importance;

[0014] According to the mapping relationship table, the first feature data and the second feature data are processed respectively to obtain a dimension judgment matrix corresponding to the first feature data and a feature judgment matrix corresponding to the second feature data.

[0015] Furthermore, the step of processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user includes:

[0016] Performing consistency check on the dimension judgment matrix and the feature judgment matrix, and setting the dimension judgment matrix and the feature judgment matrix that pass the check as the target judgment matrix;

[0017] Determining the importance features of the user feature data according to the target judgment matrix, wherein the importance features represent the importance of each user feature data to the score;

[0018] The user data of each candidate user corresponding to the target activity is processed according to the importance feature to obtain the first user score.

[0019] Furthermore, the consistency check of the dimension judgment matrix and the feature judgment matrix includes:

[0020] Obtaining a first verification index corresponding to the dimension judgment matrix and a second verification index corresponding to the feature judgment matrix;

[0021] Determining, according to the first verification index and the second verification index, a dimension consistency ratio corresponding to the dimension judgment matrix and a feature consistency ratio corresponding to the feature judgment matrix;

[0022] Based on the dimension consistency ratio and the feature consistency ratio, a consistency check is performed on the dimension judgment matrix and the feature judgment matrix.

[0023] Furthermore, the processing of the user data of each candidate user corresponding to the target activity according to the importance feature to obtain the first user score includes:

[0024] Acquire user data of each candidate user corresponding to the target activity, and construct a target data set corresponding to the user data;

[0025] The importance feature is fused with the target data set to obtain the first user score.

[0026] Furthermore, the constructing of a target data set corresponding to the user data includes:

[0027] Arranging the user data according to the user characteristic data to obtain an initial data set;

[0028] Performing maximum value conversion on the user data in the initial data set, and setting the initial data set after the maximum value conversion as the target data set.

[0029] Furthermore, the obtaining of the information entropy weight of each candidate user includes:

[0030] Normalizing the user data of the candidate user to obtain normalized data;

[0031] The total amount of normalized data belonging to the same user characteristic data is counted, and the information entropy weight is determined according to the size relationship between the normalized data and the total amount of data.

[0032] Accordingly, an embodiment of the present application provides a user rating device, comprising:

[0033] A classification module, used to obtain user feature dimensions of a target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions;

[0034] A first processing module is used to input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively;

[0035] A second processing module is used to process the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user;

[0036] An evaluation module, used to obtain the information entropy weight of each candidate user, evaluate the user data of each candidate user according to the information entropy weight, and obtain a second user score of each candidate user;

[0037] A push module is used to determine a target user according to the first user score and the second user score of each candidate user, and push the target activity to the target user.

[0038] In addition, an embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the user rating method provided in the embodiment of the present application.

[0039] In addition, an embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. The computer program is suitable for loading by a processor to execute any user rating method provided in the embodiment of the present application.

[0040] In addition, an embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements any user rating method provided in the embodiment of the present application.

[0041] The present application obtains the user feature dimensions of the target activity, classifies the user feature dimensions, obtains the user feature data corresponding to each user feature dimension, and realizes the accurate acquisition of the user feature dimensions and the user feature data respectively; then, each user feature dimension and its user feature data are input into the preset hierarchical analysis model for processing, and the dimension judgment matrix and the feature judgment matrix are obtained, which are judgment matrices corresponding to the user feature dimensions and the user feature data respectively; then, the user data of each candidate user corresponding to the target activity are processed through the dimension judgment matrix and the feature judgment matrix to obtain the first user score of each candidate user; the information entropy weight of each candidate user is obtained, the user data of each candidate user is evaluated according to the information entropy weight, and the second user score of each candidate user is obtained, and the first user score and the second user score of the candidate user are accurately calculated, so that the user can be globally scored through the first user score and the second user score; finally, according to the first user score and the second user score of each candidate user, the target user is determined and the target activity is pushed to the target user, and finally the evaluation of the global user is realized to the greatest extent, and different weights can be automatically assigned to the features for fusion, avoiding the use of invalid data caused by subjective human factors, and further improving the accuracy of user evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 It is a flowchart of a user rating method provided in an embodiment of the present application;

[0044] Figure 2 is a schematic diagram of the structure of a user rating device provided in an embodiment of the present application;

[0045] Figure 3 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0047] The embodiment of the present application provides a user rating method, device, electronic device and computer storage medium. The user rating device can be integrated in an electronic device, which can be a server or a terminal.

[0048] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), as well as big data and artificial intelligence platforms.

[0049] The terminal may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, and this application does not limit this.

[0050] In addition, the term "plurality" in the embodiments of the present application refers to two or more than two. The terms "first" and "second" in the embodiments of the present application are used to distinguish descriptions and should not be understood as implying relative importance.

[0051] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments.

[0052] See also Figure 1 , Figure 1 : is a flowchart of a user rating method provided by an embodiment of the present application. The user rating method may include:

[0053] S101 : Acquire user feature dimensions of a target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions.

[0054] In this embodiment, the target activity is the implementation activity corresponding to the target object of current concern, for example, the implementation activities corresponding to the target objects such as pulling new users, retaining users or recalling users. Different target activities correspond to different user feature dimensions. Taking the implementation activities corresponding to the retained users in the logistics scenario as an example, the user feature dimension includes the user basic piece quantity information class, the user access behavior class and the user coupon usage class. By classifying the user feature dimension, the user feature data under each different user feature dimension can be determined. For example, the user feature data under the basic piece quantity information class includes: the piece quantity of the month, the number of days since the last order, the standard deviation of the piece quantity in the previous 6 months, the fitted slope of the piece quantity in the previous 6 months, and the kurtosis and skewness of the piece quantity in the previous month, a total of 5 user feature data; the user feature data under the user access behavior class includes: the number of visits to the activity page, the total number of clicks, the number of days since the mini program was launched in the previous 7 days, and the number of days since the last launch, a total of 4 user feature data; the user feature data under the user coupon usage class includes: the number of coupons in the current account, the current account minimum discount, the current account stored value balance and the average coupon usage rate in the previous 3 months, a total of 4 user feature data. From this, we can see that the user characteristic dimension of the target activity corresponding to retained users in the logistics scenario includes a total of 13 user characteristic data.

[0055] S102: Input each of the user feature dimensions and the user feature data into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data, respectively.

[0056] In this embodiment, the hierarchical analysis model is an unsupervised model, including a target layer, a criterion layer and a scheme layer. The target layer is the target activity confirmation layer, the criterion layer is the user feature dimension and the user feature data acquisition layer, and the scheme layer is the analysis layer. Obtain a preset hierarchical analysis model, and process the user feature dimension and the user feature data respectively according to the scheme layer of the hierarchical analysis model to obtain a dimension judgment matrix and a feature judgment matrix. Specifically, the user feature dimension and the user feature data are respectively scaled in importance to obtain the first feature data and the second feature data. Among them, different user feature dimensions and different user feature data have different importance relative to the target activity. According to the expert team, each user feature dimension and each user feature data can be scaled in importance relative to the target activity, thereby obtaining the first feature data corresponding to the user feature dimension and the second feature data corresponding to the user feature data. Then, the feature data in the first feature data are compared in pairs to obtain a dimension judgment matrix; the feature data in the second feature data are compared in pairs according to the user feature dimension to which they belong, and the feature judgment matrix corresponding to the user feature data of each user feature dimension is obtained. The judgment matrix corresponding to the user feature dimension is the dimension judgment matrix, and the judgment matrix corresponding to the user feature data is the feature judgment matrix.

[0057] Taking the three user feature dimensions in the above logistics scenario as an example, the feature data in the first feature data corresponding to the user feature dimension are compared to obtain a 3*3 dimension judgment matrix; according to the user feature dimension, the second feature data corresponding to the user feature data under the same user feature dimension are compared in pairs to obtain the feature judgment matrix corresponding to each user feature data. Among them, the basic piece quantity information category includes 5 user feature data, and the second feature data corresponding to the 5 user feature data are compared in pairs to obtain a 5*5 feature judgment matrix; the user access behavior category includes 4 user feature data, and the feature data of the second feature data corresponding to the 4 user feature data are compared in pairs to obtain a 4*4 feature judgment matrix; the user coupon usage category includes 4 user feature data, and the feature data of the second feature data corresponding to the 4 user feature data are compared in pairs to obtain a 4*4 feature judgment matrix. In this way, a dimension judgment matrix and 3 feature judgment matrices are finally obtained.

[0058] S103: Process the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user.

[0059] In this embodiment, when the dimension judgment matrix and the feature judgment matrix are obtained, the user data of the candidate users corresponding to the target activity are obtained. The user data includes the user ID, and the index values ​​of each user ID stored under different user feature data, such as the consumption value and behavior value of each user ID. The user data is processed according to the dimension judgment matrix and the feature judgment matrix to obtain the first user score corresponding to each candidate user. Specifically, when the user data of each candidate user is obtained, a data frame or a user matrix corresponding to all candidate users can be constructed through the user data; and the feature weights are determined according to the dimension judgment matrix and the feature judgment matrix, and the data frame or the user matrix is ​​evaluated according to the feature weights to obtain the first user score.

[0060] S104: Obtain an information entropy weight of each candidate user, and evaluate user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user.

[0061] In this embodiment, the information entropy weight is calculated for the user feature data by the entropy weight method to obtain the weight of the user feature data again. Specifically, when the user data of each candidate user is obtained, the user data is normalized to obtain normalized data; then, the information entropy weight is calculated based on the normalized data, and the information entropy weight is multiplied by the data frame or user matrix constructed by all user data to obtain the second user score corresponding to each candidate user.

[0062] S105: Determine a target user according to the first user score and the second user score of each candidate user, and push the target activity to the target user.

[0063] In this embodiment, when the first user score and the second user score are obtained, the first user score and the second user score are merged to obtain a total score. Specifically, the preset weights corresponding to the first user score and the second user score are obtained, and the first user score and the second user score are weighted averaged according to the preset weights to calculate the total score. In addition, the first user score and the second user score can be calculated according to the distribution of the information entropy weights calculated by the entropy weight method to obtain the total score. Then, the candidate users are sorted and screened according to the total score, and the candidate users whose total scores are within the preset range are determined as target users, and the target activities are pushed to the target users.

[0064] Furthermore, since the hierarchical analysis model in this embodiment is an unsupervised model, the effect of the unsupervised model cannot be measured by a dedicated indicator. Therefore, before calculating the user score based on the hierarchical analysis model, in the process of training the initial model through the training sample, the user feature data can be clustered, and the clustering result and the initial model calculation result ranking can be observed to be consistent, thereby determining whether the initial model training is completed. If the ranking of the clustering result and the scoring result are consistent, it is determined that the initial model training is completed, and the trained initial model is the hierarchical analysis model; if the ranking of the clustering result and the scoring result are inconsistent, it is determined that the initial model training is not completed, and the parameters of the initial model are continued to be adjusted, and the training sample is continued to be invested to train the initial model after the parameter adjustment until the ranking of the clustering result and the scoring result are consistent. In addition, after the candidate users are screened by the total score, the recommendation of target activities such as issuing coupons can be made for the screened target users, and the feedback of information such as the number of times the target users use the coupons can be used to determine whether the current hierarchical analysis model needs to be further trained and improved.

[0065] This embodiment obtains the user feature dimensions of the target activity, classifies the user feature dimensions, and obtains the user feature data corresponding to each user feature dimension; then, each user feature dimension and its user feature data are input into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, which are judgment matrices corresponding to the user feature dimensions and the user feature data, respectively; then, the user data of each candidate user corresponding to the target activity is processed through the dimension judgment matrix and the feature judgment matrix to obtain the first user score of each candidate user; the information entropy weight of each candidate user is obtained, and the user data of each candidate user is evaluated according to the information entropy weight to obtain the second user score of each candidate user; finally, according to the first user score and the second user score of each candidate user, the target user is determined and the target activity is pushed to the target user, thereby achieving the maximum evaluation of global users, and being able to automatically assign different weights to features for fusion, avoiding the use of invalid data caused by subjective human intervention, and further improving the accuracy of user evaluation.

[0066] In some embodiments of the present application, the user feature dimensions and the user feature data are input into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, including:

[0067] Based on the hierarchical analysis model, the user feature dimension and the user feature data are processed respectively to obtain first feature data corresponding to the user feature dimension and second feature data corresponding to the user feature data;

[0068] Obtain a mapping relationship table of preset feature data and importance;

[0069] According to the mapping relationship table, the first feature data and the second feature data are processed respectively to obtain a dimension judgment matrix corresponding to the first feature data and a feature judgment matrix corresponding to the second feature data.

[0070] In this embodiment, the first feature data is feature data obtained by importance scaling the user feature dimension, and the second feature data is feature data obtained by importance scaling the user feature data. Importance scaling means that when the user feature dimension and the user feature data are obtained, the importance of the user feature dimension relative to the target activity is evaluated through the hierarchical analysis model to obtain the first feature data; the importance of the user feature data relative to the target activity is evaluated to obtain the second feature data. The preset mapping relationship table is a mapping table of feature data and importance, wherein the mapping scale in the mapping relationship table represents the importance of one feature data compared to another feature data. If the two feature data are the same, the mapping scale of the two feature data in the mapping relationship table is determined to be 1; if one of the two feature data is greater than the other feature data, and the difference is within the first preset range, the mapping scale corresponding to the two feature data is determined to be 3; if one of the two feature data is greater than the other feature data, and the difference is within the second preset range, the mapping scale corresponding to the two feature data is determined to be 5; if one of the two feature data is greater than the other feature data, and the difference is within the third preset range, the mapping scale corresponding to the two feature data is determined to be 7; if one of the two feature data is greater than the other feature data, and the difference is within the fourth preset range, the mapping scale corresponding to the two feature data is determined to be 9. The first preset range, the second preset range, the third preset range, and the fourth preset range are increased in sequence, and the larger the range, the larger the difference between the two feature data, and the mapping scale is not limited to the numerical value in this embodiment.

[0071] The first feature data are compared in pairs, and the mapping scale corresponding to each comparison result is determined according to the mapping relationship table, so as to obtain a dimension judgment matrix; the second feature data are compared in pairs, and the mapping scale corresponding to each comparison result is determined according to the mapping relationship table, so as to obtain a dimension judgment matrix. When comparing the second feature data, the comparison is performed according to the user feature dimension to which each second feature data belongs, that is, the second feature data under the same user dimension are compared in pairs, and the feature judgment matrix corresponding to each different user feature dimension is obtained; the second feature data under different user feature dimensions are not compared.

[0072] This embodiment compares the feature data of the first feature data and the second feature data respectively through a mapping relationship table to obtain a dimension judgment matrix and a feature judgment matrix, so that the importance of the user in different dimensions can be accurately judged through the two judgment matrices, and then the user score can be accurately calculated based on the judgment matrix.

[0073] In some embodiments of the present application, the above-mentioned processing of the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain the first user score of each candidate user includes:

[0074] Performing consistency check on the dimension judgment matrix and the feature judgment matrix, and setting the dimension judgment matrix and the feature judgment matrix that pass the check as the target judgment matrix;

[0075] Determining the importance features of the user feature data according to the target judgment matrix, wherein the importance features represent the importance of each user feature data to the score;

[0076] The user data of each candidate user corresponding to the target activity is processed according to the importance feature to obtain the first user score.

[0077] In this embodiment, in order to evaluate the rationality of the established judgment matrix, it is necessary to perform consistency checks on the dimension judgment matrix and the feature judgment matrix respectively. Specifically, the dimension consistency ratio and feature consistency ratio of the dimension judgment matrix and the feature judgment matrix are calculated respectively. The dimension consistency ratio is the consistency ratio calculated according to the dimension judgment matrix, and the feature consistency ratio is the consistency ratio calculated according to the feature judgment matrix. Determine whether the dimension consistency ratio is less than or equal to the first judgment threshold. If the dimension consistency ratio is less than or equal to the first judgment threshold, it is determined that the consistency verification of the dimension judgment matrix has passed, and the dimension judgment matrix that has passed the verification is the target judgment matrix. Determine whether the feature consistency ratio is less than or equal to the second judgment threshold. If the feature consistency ratio is less than or equal to the second judgment threshold, it is determined that the consistency verification of the feature judgment matrix has passed, and the feature judgment matrix that has passed the verification is also the target judgment matrix.

[0078] The importance feature is the importance of each user feature data to the score. When the target judgment matrix is ​​obtained, the maximum eigenvalue of the target judgment matrix is ​​calculated. Specifically, taking 4 target judgment matrices as an example, the 4 target judgment matrices include 1 target judgment matrix corresponding to 1 dimension judgment matrix and 3 target judgment matrices corresponding to 3 feature judgment matrices. The maximum eigenvalue of each target judgment matrix is ​​calculated respectively. The eigenvector corresponding to each maximum eigenvalue is determined, that is, the feature weight corresponding to each target judgment matrix is ​​obtained. In the above logistics scenario example, the feature weights corresponding to the 4 target judgment matrices can be calculated as W0, W1, W2 and W3, respectively, where W0 is a 1*3 vector, W1 is a 1*5 vector, W2 and W3 are 1*4 vectors, W0 is the feature weight corresponding to the dimension judgment matrix, and W1, W2 and W3 are the feature weights corresponding to the 3 feature judgment matrices. The feature weights corresponding to the dimension judgment matrix are multiplied by the feature weights corresponding to the feature judgment matrix to obtain the importance features corresponding to the user feature data. In the above example, by multiplying the first, second and third elements in W0 with W1, W2 and W3 respectively, the importance features of 13 user feature data can be calculated.

[0079] After obtaining the importance feature, the user data of each candidate user corresponding to the target activity is processed according to the importance feature to obtain the first user score. Specifically, the user data is constructed to obtain a target data set corresponding to the user data, the target data set is multiplied by the importance feature to obtain an evaluation result, and then the values ​​in the evaluation result are added row by row to obtain the first user score corresponding to each candidate user.

[0080] This embodiment processes the user data of each candidate user through the importance feature to obtain the first user score, thereby achieving accurate and efficient calculation of the first user score, improving the accuracy of user evaluation, further improving the accuracy of user screening, and making the recommended target activities more in line with the screened target users.

[0081] In some embodiments of the present application, the consistency check of the dimension judgment matrix and the feature judgment matrix includes:

[0082] Obtaining a first verification index corresponding to the dimension judgment matrix and a second verification index corresponding to the feature judgment matrix;

[0083] Determining, according to the first verification index and the second verification index, a dimension consistency ratio corresponding to the dimension judgment matrix and a feature consistency ratio corresponding to the feature judgment matrix;

[0084] Based on the dimension consistency ratio and the feature consistency ratio, a consistency check is performed on the dimension judgment matrix and the feature judgment matrix.

[0085] In this embodiment, the judgment matrix can be consistency checked by consistency index and consistency ratio. Specifically, the first check index is a check index randomly generated for the dimension judgment matrix, and the second check index is a check index randomly generated for the feature judgment matrix. The first check index and the second check index can be obtained by randomly generating a standard consistency comparison matrix and then calculating the average eigenvector of the standard consistency comparison matrix. In addition, the first check index and the second check index can use the same random index or different random indicators.

[0086] Afterwards, the dimension consistency ratio corresponding to the dimension judgment matrix and the feature consistency ratio corresponding to the feature judgment matrix are determined according to the first verification index and the second verification index, respectively. Specifically, the maximum eigenvalue of the eigenvector of the dimension judgment matrix is ​​calculated, and the dimension index corresponding to the dimension judgment matrix is ​​calculated according to the order of the dimension judgment matrix and the maximum eigenvalue; the feature index corresponding to the feature judgment matrix is ​​calculated according to the maximum eigenvalue of the eigenvector of the feature judgment matrix and the order of the feature judgment matrix. When the dimension index and the feature index are calculated, the ratio of the dimension index and the first verification index is calculated to obtain the dimension consistency ratio; the ratio of the feature index and the second verification index is calculated to obtain the feature consistency ratio. Then, the preset first judgment threshold and the second judgment threshold are obtained, wherein the first judgment threshold can be the same as the second judgment threshold, such as 0.1. It is determined that the consistency check of the dimension judgment matrix whose dimension consistency ratio is less than or equal to the first judgment threshold passes; it is determined that the consistency check of the feature judgment matrix whose feature consistency ratio is less than or equal to the second judgment threshold passes.

[0087] This embodiment randomly generates a first verification indicator and a second verification indicator, and calculates a consistency ratio based on the second verification indicator and the feature indicator, thereby achieving accurate verification of the consistency of the feature judgment matrix and the dimension judgment matrix, making the obtained target judgment matrix more reasonable, and further improving the accuracy of user evaluation.

[0088] In some embodiments of the present application, the processing of the user data of each candidate user corresponding to the target activity according to the importance feature to obtain the first user score includes:

[0089] Acquire user data of each candidate user corresponding to the target activity, and construct a target data set corresponding to the user data;

[0090] The importance feature is fused with the target data set to obtain the first user score.

[0091] In this embodiment, the candidate users corresponding to the target activity can be all users stored in the current database, and the user data of the candidate users can be obtained by querying the activity information corresponding to each user ID in the database. Specifically, the user data of the candidate users includes the user ID of each candidate user, and the index value of each user ID under different user feature data. For example, the user feature data is the number of visits to the activity page, and the index value under the user feature data is the number of times the candidate user visits the activity page. When all user data are obtained, a target data set corresponding to all user data is constructed. The target data set can be obtained by constructing a data frame or user matrix corresponding to the user data. The data frame is a data structure of the data set. Then, the target data set is fused with the importance feature to obtain the first user score. Among them, the target data set and the importance feature can be fused by matrix multiplication or addition. By fusing the target data set with the importance feature, the importance feature can be assigned to each data in the target data set, thereby obtaining the first user score.

[0092] This embodiment constructs a target data set corresponding to the user data, fuses the importance feature with the target data set, obtains the first user score, and achieves accurate acquisition of the first user score.

[0093] In some embodiments of the present application, the above-mentioned construction of the target data set corresponding to the user data includes:

[0094] Arranging the user data according to the user characteristic data to obtain an initial data set;

[0095] Performing maximum value conversion on the user data in the initial data set, and setting the initial data set after the maximum value conversion as the target data set.

[0096] In this embodiment, when the user data of all candidate users are obtained, the user data of each candidate user is arranged according to the user feature data to obtain an initial data set. Specifically, the user ID of the candidate user is used as the first column, one row corresponds to one candidate user, one column corresponds to one user feature data, and starting from the second column, the index value of each candidate user under the corresponding user feature data is arranged according to the user feature data to obtain the initial data set. For example, for 400,000 candidate users, under the above 13 user feature data, the initial data set of 400,000 users is composed of 400,000*14 user feature data, where the first column is the user ID of the 400,000 candidate users, and the 13 columns are the index values ​​of each candidate user under the above 13 user feature data. The user data in the initial data set is converted to a maximum value, and the initial data set after the maximum value conversion is set as the target data set.

[0097] This embodiment constructs a target data set corresponding to user data so that candidate users can be accurately scored through the target data set, thereby improving user evaluation accuracy and evaluation efficiency.

[0098] In some embodiments of the present application, the step of obtaining the information entropy weight of each candidate user includes:

[0099] Normalizing the user data of the candidate user to obtain normalized data;

[0100] The total amount of normalized data belonging to the same user characteristic data is counted, and the information entropy weight is determined according to the size relationship between the normalized data and the total amount of data.

[0101] In this embodiment, when determining the second user score of the candidate user, the index value of each candidate user under the user feature data is obtained, and the index value is normalized to obtain normalized data. The normalized data can eliminate the influence of different index dimensions and units. Taking the above 13 user feature data as an example, the index value of each candidate user under the 13 user feature data is normalized, that is, the normalized data corresponding to each candidate user under the 13 user feature data is obtained. The total amount of normalized data under the same user feature data is counted, and the information entropy weight of the candidate user is determined according to the size relationship between the normalized data and the total amount of data. Specifically, the total amount of normalized data under the same user feature data is counted, and the proportion of the normalized data in the total amount of data under the user feature data to which it belongs is calculated, and the proportion reflects the size relationship between the normalized data and the total amount of data. Then, the information entropy value of each normalized data is calculated according to the proportion, and the purity and uncertainty of the index value under each user feature data are measured by the information entropy value. Finally, the information entropy weight of each normalized data is calculated according to the information entropy value, and the information entropy weight is multiplied by the target data set corresponding to the candidate user to obtain the second user score.

[0102] This embodiment determines the information entropy weight of each candidate user, so that the candidate users can be scored twice according to the information entropy weight, avoiding inaccurate user evaluation caused by a single score, and further improving the accuracy of user evaluation.

[0103] In order to better implement the user rating method provided in the embodiment of the present application, the embodiment of the present application also provides a device based on the above user rating method. The meanings of the terms are the same as those in the above user rating method, and the specific implementation details can refer to the description in the method embodiment.

[0104] For example, Figure 2As shown, the user scoring device may include: a classification module 201, a first processing module 202, a second processing module 203, an evaluation module 204 and a push module 205.

[0105] The classification module 201 is used to obtain user feature dimensions of the target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions;

[0106] A first processing module 202 is used to input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively;

[0107] In one embodiment of the present application, the first processing module 202 includes:

[0108] A first processing unit, configured to process the user feature dimension and the user feature data respectively based on the hierarchical analysis model to obtain first feature data corresponding to the user feature dimension and second feature data corresponding to the user feature data;

[0109] A first acquisition unit, used to acquire a mapping relationship table of preset feature data and importance;

[0110] The second processing unit is used to process the first feature data and the second feature data respectively according to the mapping relationship table to obtain a dimension judgment matrix corresponding to the first feature data and a feature judgment matrix corresponding to the second feature data.

[0111] The second processing module 203 is used to process the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user;

[0112] In one embodiment of the present application, the second processing module 203 includes:

[0113] A first verification unit is used to perform consistency verification on the dimension judgment matrix and the feature judgment matrix, and set the dimension judgment matrix and the feature judgment matrix that pass the verification as a target judgment matrix;

[0114] A first confirmation unit, configured to determine the importance feature of the user feature data according to the target judgment matrix, wherein the importance feature represents the importance of each user feature data to the score;

[0115] The third processing unit is used to process the user data of each candidate user corresponding to the target activity according to the importance feature to obtain the first user score.

[0116] In one embodiment of the present application, the first verification unit includes:

[0117] A second acquisition unit, used to acquire a first verification indicator corresponding to the dimension judgment matrix and a second verification indicator corresponding to the feature judgment matrix;

[0118] A second confirmation unit is used to determine the dimension consistency ratio corresponding to the dimension judgment matrix and the feature consistency ratio corresponding to the feature judgment matrix according to the first verification index and the second verification index respectively;

[0119] The second verification unit is used to perform consistency verification on the dimension judgment matrix and the feature judgment matrix based on the dimension consistency ratio and the feature consistency ratio.

[0120] In one embodiment of the present application, the third processing unit includes:

[0121] A construction unit, configured to obtain user data of each candidate user corresponding to the target activity, and construct a target data set corresponding to the user data;

[0122] A fusion unit is used to fuse the importance feature with the target data set to obtain the first user score.

[0123] In one embodiment of the present application, the above-mentioned building unit includes:

[0124] an arranging unit, configured to arrange the user data according to the user characteristic data to obtain an initial data set;

[0125] The conversion unit is used to perform maximum value conversion on the user data in the initial data set, and set the initial data set after the maximum value conversion as the target data set.

[0126] An evaluation module 204 is used to obtain an information entropy weight of each candidate user, and evaluate the user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user;

[0127] In one embodiment of the present application, the evaluation module 204 includes:

[0128] A normalization unit, configured to perform normalization processing on the user data of the candidate user to obtain normalized data;

[0129] The statistical unit is used to count the total amount of normalized data belonging to the same user characteristic data, and determine the information entropy weight according to the size relationship between the normalized data and the total amount of data.

[0130] The push module 205 is configured to determine a target user according to the first user score and the second user score of each candidate user, and push the target activity to the target user.

[0131] In specific implementation, the above modules can be implemented as independent entities, or can be arbitrarily combined and implemented as the same or several entities. The specific implementation methods and corresponding beneficial effects of the above modules can be found in the previous method embodiments, which will not be repeated here.

[0132] The present application also provides an electronic device, which may be a server or a terminal, etc. Figure 3 As shown, it shows a schematic diagram of the structure of the electronic device involved in the embodiment of the present application, specifically:

[0133] The electronic device may include components such as a processor 601 with one or more processing cores, a memory 602 with one or more computer-readable storage media, a power supply 603, and an input unit 604. Those skilled in the art will appreciate that Figure 3 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0134] The processor 601 is the control center of the electronic device, which uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing computer programs and / or modules stored in the memory 602, and calling data stored in the memory 602. Optionally, the processor 601 may include one or more processing cores; preferably, the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 601.

[0135] The memory 602 can be used to store computer programs and modules. The processor 601 executes various functional applications and data processing by running the computer programs and modules stored in the memory 602. The memory 602 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, a computer program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 602 may also include a memory controller to provide the processor 601 with access to the memory 602.

[0136] The electronic device also includes a power supply 603 for supplying power to each component. Preferably, the power supply 603 can be logically connected to the processor 601 through a power management system, so as to manage charging, discharging, power consumption and other functions through the power management system. The power supply 603 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0137] The electronic device may further include an input unit 604, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0138] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 601 in the electronic device will load the executable files corresponding to the processes of one or more computer programs into the memory 602 according to the following instructions, and the processor 601 will run the computer programs stored in the memory 602, thereby realizing various functions, such as:

[0139] Acquire user feature dimensions of the target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions;

[0140] Input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively;

[0141] Processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user;

[0142] Obtaining an information entropy weight of each candidate user, and evaluating user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user;

[0143] A target user is determined according to the first user score and the second user score of each candidate user, and the target activity is pushed to the target user.

[0144] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the detailed description of the user rating method above, which will not be elaborated here.

[0145] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by a computer program, or by controlling related hardware through a computer program. The computer program may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0146] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, and the computer program can be loaded by a processor to execute the steps in any user rating method provided in the embodiment of the present application. For example, the computer program can execute the following steps:

[0147] Acquire user feature dimensions of the target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions;

[0148] Input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively;

[0149] Processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user;

[0150] Obtaining an information entropy weight of each candidate user, and evaluating user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user;

[0151] A target user is determined according to the first user score and the second user score of each candidate user, and the target activity is pushed to the target user.

[0152] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the previous embodiments and will not be described in detail here.

[0153] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0154] Since the computer program stored in the computer-readable storage medium can execute the steps in any one of the user rating methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the user rating methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0155] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above-mentioned user rating method.

[0156] The above is a detailed introduction to a user rating method, device, electronic device and computer storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A user rating method, characterized in that: include: Acquire user feature dimensions of the target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions; Input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively; Processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user; Obtaining an information entropy weight of each candidate user, and evaluating user data of each candidate user according to the information entropy weight to obtain a second user score of each candidate user; A target user is determined according to the first user score and the second user score of each candidate user, and the target activity is pushed to the target user.

2. The user rating method according to claim 1, characterized in that: The step of inputting each of the user feature dimensions and the user feature data into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix includes: Based on the hierarchical analysis model, the user feature dimension and the user feature data are processed respectively to obtain first feature data corresponding to the user feature dimension and second feature data corresponding to the user feature data; Obtain a mapping relationship table of preset feature data and importance; According to the mapping relationship table, the first feature data and the second feature data are processed respectively to obtain a dimension judgment matrix corresponding to the first feature data and a feature judgment matrix corresponding to the second feature data.

3. The user rating method according to claim 1, characterized in that: The step of processing the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user includes: Performing consistency check on the dimension judgment matrix and the feature judgment matrix, and setting the dimension judgment matrix and the feature judgment matrix that pass the check as the target judgment matrix; Determining the importance features of the user feature data according to the target judgment matrix, wherein the importance features represent the importance of each user feature data to the score; The user data of each candidate user corresponding to the target activity is processed according to the importance feature to obtain the first user score.

4. The user rating method according to claim 3, characterized in that: The performing consistency check on the dimension judgment matrix and the feature judgment matrix includes: Obtaining a first verification index corresponding to the dimension judgment matrix and a second verification index corresponding to the feature judgment matrix; Determining a dimension consistency ratio corresponding to the dimension judgment matrix and a feature consistency ratio corresponding to the feature judgment matrix according to the first verification index and the second verification index respectively; Based on the dimension consistency ratio and the feature consistency ratio, a consistency check is performed on the dimension judgment matrix and the feature judgment matrix.

5. The user rating method according to claim 3, characterized in that: The step of processing the user data of each candidate user corresponding to the target activity according to the importance feature to obtain the first user score includes: Acquire user data of each candidate user corresponding to the target activity, and construct a target data set corresponding to the user data; The importance feature is fused with the target data set to obtain the first user score.

6. The user rating method according to claim 5, characterized in that: The constructing a target data set corresponding to the user data includes: Arranging the user data according to the user characteristic data to obtain an initial data set; Performing maximum value conversion on the user data in the initial data set, and setting the initial data set after the maximum value conversion as the target data set.

7. The user rating method according to claim 1, characterized in that: The obtaining of the information entropy weight of each candidate user includes: Normalizing the user data of the candidate user to obtain normalized data; The total amount of normalized data belonging to the same user characteristic data is counted, and the information entropy weight is determined according to the size relationship between the normalized data and the total amount of data.

8. A user rating device, characterized in that: include: A classification module, used to obtain user feature dimensions of a target activity, classify the user feature dimensions, and obtain user feature data corresponding to each of the user feature dimensions; A first processing module is used to input each of the user feature dimensions and the user feature data thereof into a preset hierarchical analysis model for processing to obtain a dimension judgment matrix and a feature judgment matrix, wherein the dimension judgment matrix and the feature judgment matrix are judgment matrices corresponding to the user feature dimensions and the user feature data respectively; A second processing module is used to process the user data of each candidate user corresponding to the target activity through the dimension judgment matrix and the feature judgment matrix to obtain a first user score of each candidate user; An evaluation module, used to obtain the information entropy weight of each candidate user, evaluate the user data of each candidate user according to the information entropy weight, and obtain a second user score of each candidate user; A push module is used to determine a target user according to the first user score and the second user score of each candidate user, and push the target activity to the target user.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the steps in the user rating method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the steps in the user rating method according to any one of claims 1 to 7.

Images