CN110598949A - User interest degree analysis method and device, electronic equipment and storage medium - Google Patents
User interest degree analysis method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN110598949A CN110598949A CN201910894901.2A CN201910894901A CN110598949A CN 110598949 A CN110598949 A CN 110598949A CN 201910894901 A CN201910894901 A CN 201910894901A CN 110598949 A CN110598949 A CN 110598949A
- Authority
- CN
- China
- Prior art keywords
- behavior data
- user
- feature vector
- matrix
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a user interest degree analysis method, which comprises the following steps: acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data; and generating a multidimensional feature vector of the target user for the target object based on the user behavior data, and processing the multidimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the target user's interest degree in the target object. Based on the technical scheme disclosed by the invention, the interest degree of the user on the target object is accurately determined.
Description
Technical Field
The invention belongs to the technical field of internet, and particularly relates to a user interest degree analysis method and device, electronic equipment and a storage medium.
Background
In internet products, recommendation systems are widely used. The recommendation system generally determines or predicts the preference or interest of the user based on big data and algorithm, so as to recommend the object which is as good as possible to the preference or interest of the user, and improve the recommendation success rate.
If the analysis for the user interest level is deviated, the deviation between the recommended content to the user and the actual demand of the user is inevitably caused, and the resource utilization rate is reduced.
Disclosure of Invention
In view of this, the present invention provides a method and an apparatus for analyzing user interest, an electronic device, and a storage medium, so as to accurately determine the interest of a user in a target object, thereby providing a more accurate recommendation basis for a recommendation system, and making content recommended to the user by the recommendation system closer to the user's requirement.
In order to achieve the purpose, the invention provides the following technical scheme:
in one aspect, the present invention provides a method for analyzing user interest, including:
acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
generating a multi-dimensional feature vector for the target user for the target object based on the user behavior data, each dimension of the multi-dimensional feature vector corresponding to one behavior data of the user behavior data;
processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
Optionally, on the basis of the above method, the method further includes: predetermining a feature vector weight matrix; wherein the process of predetermining the eigenvector weight matrix comprises:
obtaining a sample behavior data set, wherein the sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data;
generating a sample feature vector matrix based on the sample behavior dataset;
calculating a covariance matrix of the sample eigenvector matrix;
solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm;
wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0; a product of the eigenvector weight matrix and a transposed matrix of the eigenvector weight matrix is less than or equal to 1; the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequential multiplication of the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
Optionally, in the method, the solving an optimal solution of the feature vector weight matrix under the condition that the optimal solution meets the preset constraint condition by using a preset algorithm includes:
and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a projection gradient descent method.
Optionally, in the above method, the obtaining user behavior data of the target user for the target object includes:
acquiring a behavior log of the target user;
determining user behavior data of the target user for the target object based on the behavior log.
In another aspect, the present invention provides a user interest level analyzing apparatus, including:
the user behavior data acquisition unit is used for acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
a multidimensional feature vector generation unit, configured to generate a multidimensional feature vector for the target object by the target user based on the user behavior data, where each dimension of the multidimensional feature vector corresponds to one behavior data of the user behavior data;
the interest prediction unit is used for processing the multi-dimensional feature vector by utilizing a predetermined feature vector weight matrix to obtain a prediction value of the interest of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
Optionally, on the basis of the above apparatus, the apparatus further includes: a weight matrix determination unit for determining a feature vector weight matrix in advance; the weight matrix determination unit includes:
the device comprises a sample behavior data set acquisition subunit, a data processing subunit and a data processing unit, wherein the sample behavior data set acquisition subunit is used for acquiring a sample behavior data set, the sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data;
a sample feature vector matrix generation subunit, configured to generate a sample feature vector matrix based on the sample behavior data set;
a covariance matrix determination subunit, configured to calculate a covariance matrix of the sample eigenvector matrix;
the solving subunit is used for solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm;
wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0; a product of the eigenvector weight matrix and a transposed matrix of the eigenvector weight matrix is less than or equal to 1; the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequential multiplication of the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
Optionally, the solving subunit solves the optimal solution of the feature vector weight matrix under the condition that the optimal solution meets the preset constraint condition by using a preset algorithm, specifically:
and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a projection gradient descent method.
Optionally, the user behavior data obtaining unit obtains user behavior data of the target user for the target object, specifically:
acquiring a behavior log of the target user; determining user behavior data of the target user for the target object based on the behavior log.
In another aspect, the present invention provides an electronic device comprising a processor and a memory;
the processor is used for calling and executing the program stored in the memory;
the memory is configured to store the program, the program at least to:
acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
generating a multi-dimensional feature vector for the target user for the target object based on the user behavior data, each dimension of the multi-dimensional feature vector corresponding to one behavior data of the user behavior data;
processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
In another aspect, the present invention provides a storage medium having stored thereon computer-executable instructions, which when loaded and executed by a processor, implement the method for analyzing user interest level as described in any one of the above.
Therefore, the beneficial effects of the invention are as follows:
the user interest degree analysis method provided by the invention comprises the steps of firstly obtaining user behavior data of a target user aiming at a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, then generating a multidimensional eigenvector of the target user aiming at the target object based on the user behavior data, and then processing the multidimensional eigenvector by utilizing a predetermined eigenvector weight matrix to obtain a predicted value of the interest degree of the target user aiming at the target object, wherein the weight of the eigenvector corresponding to the positive feedback behavior data is more than or equal to 0, and the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0.
The user interest degree analysis method provided by the invention is used for analyzing the interest degree of the user on the target object by combining the positive feedback behavior data and the negative feedback behavior data, so that more comprehensive data is provided for the user interest degree analysis; in the process of calculating the predicted value of the target object interest level of the target user, the feature vector corresponding to the positive feedback behavior data is in positive correlation with the predicted value of the interest level, and the feature vector corresponding to the negative feedback behavior data is in negative correlation with the predicted value of the interest level, so that the finally obtained predicted value of the target object interest level of the target user is closer to the real state of the user, and the accuracy is higher.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is an architecture diagram of a user interest level analysis system provided by the present invention;
FIG. 2 is a flowchart of a method for analyzing user interest according to the present invention;
FIG. 3 is a flow chart of a method for determining a feature vector weight matrix according to the present invention;
fig. 4 is a signaling diagram of a user interest analysis method in an application scenario according to the present invention;
FIG. 5 is a schematic illustration of a game aid interface for a tournament game in accordance with the present invention;
FIG. 6 is a schematic structural diagram of a user interest level analysis apparatus according to the present invention;
fig. 7 is a hardware structure diagram of an electronic device according to the present invention.
Detailed Description
The invention provides a user interest degree analysis method and device, electronic equipment and a storage medium, so that the interest degree of a user on a target object can be accurately determined, more accurate recommendation basis can be provided for a recommendation system, and the content recommended to the user by the recommendation system is closer to the requirement of the user.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The user interest degree analysis method provided by the invention can be applied to a server or a terminal. The aforementioned terminal may be an electronic device such as a desktop computer, a mobile terminal (e.g., a smart phone and a tablet computer), and the like. The aforementioned server may be one server, or a server cluster composed of a plurality of servers, or a cloud computing service center.
Referring to fig. 1, fig. 1 is an architecture diagram of a user interest level analysis system according to the present invention.
The user interest level analysis system includes a terminal 101 and a server 102. The terminal 101 and the server 102 perform data interaction through a communication network.
In one possible implementation, the terminal 101 sends user behavior data of the user for the target object to the server 102. The server 102 determines the interest degree of the user in the target object according to the user behavior data sent by the terminal 101. Optionally, the server 102 determines the content to be recommended to the terminal 101 by using the interest degree of the user in the target object as a recommendation basis.
In another possible implementation, the terminal 101 sends a behavior log to the server 102. The server 102 analyzes the behavior log sent by the terminal 101, obtains user behavior data of the user for the target object, and then determines the interest degree of the user for the target object according to the user behavior data. Optionally, the server 102 determines the content to be recommended to the terminal 101 by using the interest degree of the user in the target object as a recommendation basis.
In addition, if the server 102 does not have the content recommendation function, after determining the interest level of the user in the target object, the server 102 sends the interest level of the user in the target object to other recommendation systems, so that the recommendation systems determine the content to be recommended to the terminal 101 by taking the interest level of the user in the target object as a recommendation basis.
In another possible implementation manner, the terminal 101 determines the interest level of the user in the target object according to the user behavior data of the user for the target object, and then sends the interest level of the user in the target object to the server 102. The server 102 determines the content to be recommended to the terminal 101 by using the interest degree of the user in the target object as a recommendation basis.
It should be noted that the target object and the sample object in the present invention refer to elements capable of outputting information to a user in a page displayed by a terminal, including but not limited to videos, texts and pictures, where the texts include but not limited to articles, question and answer conversations, and game strategies.
Taking a match game installed in a terminal as an example, after the game is run, match strategies, introduction to weapons, operation skill videos and wonderful game videos displayed in a game page can be used as target objects and sample objects.
Referring to fig. 2, fig. 2 is a flowchart of a user interest analysis method provided by the present invention. The method is applicable to a server connected with a terminal, and comprises the following steps:
s201: and obtaining user behavior data of the target user aiming at the target object.
In the present invention, a user to be analyzed is referred to as a target user, and an object to be analyzed is referred to as a target object.
When a user faces a certain object, whether the object is interested or not determines the operation behavior of the user on the object.
For example, if the user is interested in a certain object, the user clicks the object, and the terminal displays the specific content of the object in response to the clicking operation. After the terminal displays the specific content of the object, the time for the user to browse the specific content of the object is longer. Furthermore, the user can make positive comments on the object, approve the object and share the object.
For example, if the user has no interest in an object, the user may not click on the object, or the user may perform a close operation shortly after clicking on the object. Further, the user may make negative comments on the object.
Therefore, in the present invention, the user behavior data includes positive feedback behavior data and negative feedback behavior data. The positive feedback behavior data represent that the user is interested in the target object, and the negative feedback behavior data represent that the user is not interested in the target object.
That is to say, the technical scheme provided by the invention combines the positive feedback behavior data and the negative feedback behavior data of the target user to the target object to comprehensively analyze the interest degree of the target user to the target object.
In practice, positive feedback behavior data includes, but is not limited to: the number of times of clicking the target object, the number of times of browsing the target object meeting preset requirements, the number of times of sharing the target object, the number of times of making a positive comment on the target object, and the number of times of making a praise on the target object. Negative feedback behavior data includes, but is not limited to: the number of times that the target object has been exposed but not clicked, the number of times that the target object is viewed that does not meet preset requirements, and the number of times that negative comments are made for the target object.
In a possible implementation manner, the aforementioned "preset requirement" is: and the time length for browsing the target object reaches a preset time threshold.
In another possible implementation manner, the aforementioned "preset requirement" is: the ratio of the target object browsing time length to the total target object browsing time length reaches a preset value. Under the condition that the target object is a video, the total browsing duration of the target object is the total duration of the video; and under the condition that the target object is a text, the total browsing time length of the target object is the estimated total time length required for reading the text.
In a possible implementation manner, the obtaining of the user behavior data of the target user for the target object specifically includes: acquiring a behavior log of a target user; and determining user behavior data of the target user for the target object based on the behavior log.
The behavior log comprises records of various operations executed by the user, and explicit behavior data and implicit behavior data of the user aiming at the target object can be determined according to the behavior log. The dominant behavior data refers to: behavior data generated by the user executing operation aiming at the target object; implicit behavior data refers to: the user does not have behavioral data resulting from performing operations on the target object. For example, the number of times of clicking on a target object, the number of times of browsing the target object meeting preset requirements, the number of times of sharing the target object, the number of times of posting positive comments on the target object, the number of times of agreeing on the target object, the number of times of browsing the target object not meeting the preset requirements, and the number of times of posting negative comments on the target object are all explicit behavior data. And the number of times the target object has been exposed but not clicked is the implicit behavior data.
S202: and generating a multi-dimensional feature vector of the target user for the target object based on the user behavior data.
Wherein each dimension in the multi-dimensional feature vector corresponds to a behavior data of the target user.
As described in the foregoing by way of example, if the user behavior data includes 5 pieces of positive feedback behavior data and 3 pieces of negative feedback behavior data, then the multidimensional feature vector generated based on the user behavior data is an 8-dimensional feature vector, which can be represented as Z1＝(z1,z2,z3,z4,z5,z6,z7,z8)。
S203: and processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object.
Wherein each dimension of the multi-dimensional eigenvector corresponds to an eigenvector weight in the eigenvector weight matrix. In addition, in the process of predetermining the eigenvector weight matrix, a preset constraint condition needs to be satisfied, and the preset constraint condition at least comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
The predicted value of the target user interest degree of the target object is obtained by processing the multidimensional characteristic vector by utilizing a predetermined characteristic vector weight matrix. The higher the predicted value is, the higher the interest degree of the target user in the target object is, and correspondingly, the lower the predicted value is, the lower the interest degree of the target user in the target object is.
It has been explained in the foregoing that positive feedback behavior data characterizes a user's interest in a target object, and negative feedback behavior data characterizes a user's disinterest in a target object. Then, in the process of calculating the predicted value of the interest level of the target user for the target object, the feature vector corresponding to the positive feedback behavior data should be in a positive correlation with the predicted value of the interest level, and the feature vector corresponding to the negative feedback behavior should be in a negative correlation with the predicted value of the interest level.
Therefore, in the process of determining the eigenvector weight matrix in advance, at least: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0. The feature vector weight corresponding to the positive feedback behavior data is greater than or equal to 0, and the feature vector weight corresponding to the negative feedback behavior data is less than or equal to 0, so that in the process of calculating the predicted value of the target user's interest level in the target object, the feature vector corresponding to the positive feedback behavior data is in positive correlation with the predicted value of the interest level, and the feature vector corresponding to the negative feedback behavior is in negative correlation with the predicted value of the interest level, so that the finally obtained predicted value of the target user's interest level in the target object is closer to the real state of the user.
It should be noted that the sample behavior data used in the process of determining the feature vector weight matrix is the same as the type of behavior data included in the user behavior data of the target user for the target object. Each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in the feature vector weight matrix.
Still in connection with the above example, the feature vector weight matrix may be represented as:
A＝(a1,a2,a3,a4,a5,a6,a7,a8)。
and (3) calculating a predicted value of the interest degree of the target user on the target object by using the formula (1).
Y＝AZ1 TFormula (1)
Wherein Y is a predicted value of the interest degree of the target user to the target object, A is a feature vector weight matrix, and Z is1 TMultidimensional feature vector Z for target user for target object1A transposed matrix (being a row matrix).
In the invention, a process of processing a multi-dimensional eigenvector by using a predetermined eigenvector weight matrix to obtain a predicted value of the target user interest degree on a target object is a process of reducing the dimension of the multi-dimensional eigenvector by using an improved principal component analysis algorithm (EPCA), can keep the maximum information quantity (namely the maximum variance of data after dimension reduction), and considers the positive influence of positive feedback behavior data on the final predicted value, the negative influence of negative feedback behavior data on the final predicted value and the influence degree of each behavior data on the final predicted value (namely each behavior data corresponds to respective weight) in the dimension reduction process.
The user interest degree analysis method provided by the invention comprises the steps of firstly obtaining user behavior data of a target user aiming at a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, then generating a multidimensional eigenvector of the target user aiming at the target object based on the user behavior data, and then processing the multidimensional eigenvector by utilizing a predetermined eigenvector weight matrix to obtain a predicted value of the interest degree of the target user aiming at the target object, wherein the weight of the eigenvector corresponding to the positive feedback behavior data is more than or equal to 0, and the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0.
The user interest degree analysis method provided by the invention is used for analyzing the interest degree of the user on the target object by combining the positive feedback behavior data and the negative feedback behavior data, so that more comprehensive data is provided for the user interest degree analysis; in the process of calculating the predicted value of the target object interest level of the target user, the feature vector corresponding to the positive feedback behavior data is in positive correlation with the predicted value of the interest level, and the feature vector corresponding to the negative feedback behavior data is in negative correlation with the predicted value of the interest level, so that the finally obtained predicted value of the target object interest level of the target user is closer to the real state of the user, and the accuracy is higher.
The following describes a process of determining the eigenvector weight matrix in advance.
Referring to fig. 3, fig. 3 is a flowchart of a method for determining a feature vector weight matrix according to the present invention. The method can also be applied to a server connected with a terminal, and comprises the following steps:
s301: a sample behavior data set is obtained.
The sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, wherein each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data. The positive feedback behavior data represents that the sample user is interested in the sample object, and the negative feedback behavior data represents that the sample user is not interested in the sample object.
Taking the number of sample users as P and the number of sample objects as Q as an example, the sample behavior data set contains P × Q sample behavior data, and each sample user contains Q sample behavior data, where P and Q are integers greater than or equal to 2.
S302: a sample feature vector matrix is generated based on the sample behavior dataset.
The sample feature vector matrix may be represented as:
wherein m is the number of sample behavior data contained in the sample behavior data set; n is the number of dimensions, consistent with the number of behavior data contained in each sample behavior data. In combination with the foregoing example, the value of m is P × Q, each sample behavior data includes 5 positive feedback behavior data and 3 negative feedback behavior data, and then the value of n is 8.
z11Characteristic value, z, representing a first dimension of the first sample behavior data12Eigenvalues of the second dimension representing the first sample behavior data, and so on, z1nA feature value representing the nth dimension of the first sample behavior data; z is a radical ofm1Characteristic value, z, of a first dimension representing the m-th sample behavior datam2Eigenvalues of the second dimension representing the m-th sample behavior data, and so on, zmnAnd representing the characteristic value of the nth dimension of the mth sample behavior data.
S303: and calculating a covariance matrix of the sample feature vector matrix.
The covariance matrix of the sample eigenvector matrix Z is denoted as X. Element X in covariance matrix X at ith row and jth columnijComprises the following steps: in the sample feature vector matrix Z, the ith column (dimension) data set ZiAnd j column (dimension) data set ZjCovariance of (c) Cov (Z)i,Zj)。
Namely:
wherein x isij＝Cov(Zi，Zj)＝(ZTZ)ij
S304: and solving the optimal solution of the feature vector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm.
Wherein the preset constraint condition comprises:
the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0;
the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0;
the product of the eigenvector weight matrix and the transposed matrix of the eigenvector weight matrix is less than or equal to 1;
and the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequentially multiplying the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
The following explains the theoretical derivation process for determining the eigenvector weight matrix:
reducing the sample feature vector matrix Z to 1 dimension, recording the matrix after dimension reduction as Y, and then:
wherein z isiRepresenting the ith row in the sample feature vector matrix Z.
Writing Y in matrix form, i.e.:
Y＝Azi T
wherein A ═ a1,a2,a3...an) Is a feature vector weight matrix; z is a radical ofi TIs the transpose of the ith row in the sample eigenvector matrix Z.
The variance of Y is:
Var(Y)＝Eγ2-(Eγ)2
＝AZTZAT
＝A·X·AT
in the process of reducing the sample feature vector matrix Z to 1 dimension, the maximum variance of Y needs to be ensured, and positive correlation relationship between some dimensions and the variance of Y needs to be defined, and negative correlation relationship between some dimensions and the variance of Y needs to be defined. The specific dimension needs to be in positive correlation with the variance of Y, and the specific dimension needs to be in negative correlation with the variance of Y, and the specific dimension needs to be determined by positive feedback behavior data or negative feedback behavior data corresponding to each dimension.
The above problem is converted into an optimization problem. The requirements are satisfied:
1) the weight of the characteristic vector corresponding to the positive feedback behavior data is greater than or equal to 0;
2) the weight of the feature vector corresponding to the negative feedback behavior data is less than or equal to 0;
3) the product of the eigenvector weight matrix and the transpose of the eigenvector weight matrix is less than or equal to 1, i.e.
In this case, the optimization target is maxvar (y) ═ a × cov (x) × aT。
In addition, since cov (x) XXTIs an orthodefinite matrix, and thus var (y) is an orthodefinite quadratic form, such optimization problems must have an optimal solution.
In one possible implementation, the optimal solution of the eigenvector weight matrix satisfying the preset constraint condition is solved by using a projection gradient descent method.
In another possible implementation manner, a preset algorithm is used to solve the optimal solution of the eigenvector weight matrix under the condition that the preset constraint condition is satisfied, specifically: and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a Newton method or a conjugate gradient method.
The following describes a user interest analysis method provided by the present invention with reference to an application scenario.
Referring to fig. 4, fig. 4 is a signaling diagram of a user interest level analysis method in an application scenario according to the present invention. The method comprises the following steps:
s401: the terminal responds to the user request, starts the battle game and displays a game interface.
S402: the terminal responds to the user request and displays the game auxiliary interface.
FIG. 5 is an example of a game aid interface. The game aid interface comprises a plurality of objects, for example: game skill video, firearm collocation skill video, firearm function introduction, game field introduction and game rule introduction.
S403: the terminal acquires user behavior data of a user aiming at a target object. For example, a firearm collocation skill video in a game assistance interface is determined to be a target object.
S404: and the terminal sends the user behavior data of the user aiming at the target object to the server.
S405: and the server determines the predicted value of the interest degree of the user on the target object according to the user behavior data.
The server determines the predicted value of the interest degree of the user in the target object based on the method shown in fig. 2.
S406: and the server determines recommended content according to the predicted value of the interest degree of the user on the target object.
S407: the server transmits the recommended content to the terminal.
S408: and the terminal displays the recommended content on the game auxiliary interface.
On the other hand, the embodiment of the invention also provides a user interest degree analysis device.
The following describes a user interest level analysis apparatus according to the present invention. The user interest analysis apparatus described below may be considered as a program module that is required to be set by the electronic device to implement the user interest analysis method provided in the embodiment of the present invention. The following description of the user interest level analyzing apparatus may be cross-referenced with the above description of the interest level analyzing method.
Referring to fig. 6, fig. 6 is a schematic structural diagram of a user interest level analysis apparatus provided in the present invention. The device includes:
a user behavior data obtaining unit 601, configured to obtain user behavior data of a target user for a target object. The user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represent that the user is interested in the target object, and the negative feedback behavior data represent that the user is not interested in the target object.
A multi-dimensional feature vector generating unit 602, configured to generate a multi-dimensional feature vector of a target user for a target object based on the user behavior data. Wherein each dimension of the multi-dimensional feature vector corresponds to one of the user behavior data.
The interestingness predicting unit 603 is configured to process the multidimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interestingness of the target user to the target object.
Each dimension of the multi-dimensional eigenvector corresponds to an eigenvector weight in the eigenvector weight matrix, and in the process of predetermining the eigenvector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
In an optional embodiment, on the basis of the user interest level analyzing device, further setting: and the weight matrix determining unit is used for determining the characteristic vector weight matrix in advance.
In an alternative embodiment, the weight matrix determining unit includes:
and the sample behavior data set acquisition subunit is used for acquiring the sample behavior data set. The sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, wherein each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data.
And the sample characteristic vector matrix generating subunit is used for generating a sample characteristic vector matrix based on the sample behavior data set.
And the covariance matrix determining subunit is used for calculating a covariance matrix of the sample feature vector matrix.
And the solving subunit is used for solving the optimal solution of the feature vector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm.
Wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0; the product of the eigenvector weight matrix and the transposed matrix of the eigenvector weight matrix is less than or equal to 1; and the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequentially multiplying the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
In a possible implementation manner, the solving subunit solves the optimal solution of the eigenvector weight matrix under the condition that the preset constraint condition is met by using a preset algorithm, specifically: and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a projection gradient descent method.
In a possible implementation manner, the user behavior data obtaining unit obtains user behavior data of the target user for the target object, specifically: acquiring a behavior log of a target user; and determining user behavior data of the target user for the target object based on the behavior log.
On the other hand, the embodiment of the invention also provides electronic equipment.
Referring to fig. 7, fig. 7 is a hardware structure diagram of an electronic device according to the present invention. The electronic device may include a processor 701 and a memory 702.
Optionally, the terminal may further include: a communication interface 703, an input unit 704, a display 705, and a communication bus 706. The processor 701, the memory 702, the communication interface 703, the input unit 704, and the display 705 all communicate with each other via the communication bus 706.
In the embodiment of the present invention, the processor 701 may be a Central Processing Unit (CPU), an application specific integrated circuit, a digital signal processor, an off-the-shelf programmable gate array or other programmable logic device.
The processor 701 may call a program stored in the memory 702.
The memory 702 is used to store one or more programs, which may include program code comprising computer operating instructions. In the embodiment of the present invention, the memory stores at least a program for realizing the following functions:
acquiring user behavior data of a target user aiming at a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
generating a multi-dimensional feature vector of a target user for a target object based on the user behavior data, wherein each dimension in the multi-dimensional feature vector corresponds to one behavior data in the user behavior data;
processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object;
each dimension of the multi-dimensional eigenvector corresponds to an eigenvector weight in the eigenvector weight matrix, and in the process of predetermining the eigenvector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
In one possible implementation, the memory 702 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, the above-mentioned programs, and the like; the storage data area may store data created during use of the computer device, and the like.
In addition, the memory 702 may include high speed random access memory, and may also include non-volatile memory.
The communication interface 703 may be an interface of a communication module.
The input unit 704 may include a touch sensing unit sensing a touch event on the touch display panel, a keyboard, and the like.
The display 705 includes a display panel, such as a touch display panel or the like.
Of course, the electronic device structure shown in fig. 7 does not limit the electronic device in the embodiment of the present invention, and in practical applications, the electronic device may include more or less components than those shown in fig. 7, or some components may be combined.
In some embodiments, the electronic device may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting a plurality of nodes through a network communication. Nodes can form a Peer-To-Peer (P2P, Peer To Peer) network, and any type of computing device, such as a server, a terminal, and other electronic devices, can become a node in the blockchain system by joining the Peer-To-Peer network.
In another aspect, the present invention further provides a storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are loaded and executed by a processor, the method for analyzing user interest level in any of the above embodiments is implemented.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device, the electronic device and the storage medium disclosed by the embodiment correspond to the method disclosed by the embodiment, so that the description is relatively simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for analyzing user interest degree, which is characterized by comprising the following steps:
acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
generating a multi-dimensional feature vector for the target user for the target object based on the user behavior data, each dimension of the multi-dimensional feature vector corresponding to one behavior data of the user behavior data;
processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition is required to be met, wherein the preset constraint condition at least comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
2. The method of claim 1, further comprising: predetermining a feature vector weight matrix; wherein the process of predetermining the eigenvector weight matrix comprises:
obtaining a sample behavior data set, wherein the sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data;
generating a sample feature vector matrix based on the sample behavior dataset;
calculating a covariance matrix of the sample eigenvector matrix;
solving the optimal solution of the feature vector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm;
wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0; a product of the eigenvector weight matrix and a transposed matrix of the eigenvector weight matrix is less than or equal to 1; the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequential multiplication of the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
3. The method according to claim 2, wherein the solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm comprises:
and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a projection gradient descent method.
4. The method of claim 1, wherein obtaining user behavior data of a target user for a target object comprises:
acquiring a behavior log of the target user;
determining user behavior data of the target user for the target object based on the behavior log.
5. A user interest level analysis apparatus, comprising:
the user behavior data acquisition unit is used for acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
a multidimensional feature vector generation unit, configured to generate a multidimensional feature vector for the target object by the target user based on the user behavior data, where each dimension of the multidimensional feature vector corresponds to one behavior data of the user behavior data;
the interest prediction unit is used for processing the multi-dimensional feature vector by utilizing a predetermined feature vector weight matrix to obtain a prediction value of the interest of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
6. The apparatus of claim 5, further comprising: a weight matrix determination unit for determining a feature vector weight matrix in advance; the weight matrix determination unit includes:
the device comprises a sample behavior data set acquisition subunit, a data processing subunit and a data processing unit, wherein the sample behavior data set acquisition subunit is used for acquiring a sample behavior data set, the sample behavior data set comprises sample behavior data of a plurality of sample users for a plurality of sample objects, each sample behavior data is sample behavior data of one sample user for one sample object, and the sample behavior data comprises positive feedback behavior data and negative feedback behavior data;
a sample feature vector matrix generation subunit, configured to generate a sample feature vector matrix based on the sample behavior data set;
a covariance matrix determination subunit, configured to calculate a covariance matrix of the sample eigenvector matrix;
the solving subunit is used for solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a preset algorithm;
wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the weight of the eigenvector corresponding to the negative feedback behavior data is less than or equal to 0; a product of the eigenvector weight matrix and a transposed matrix of the eigenvector weight matrix is less than or equal to 1; the variance of a dimensionality reduction matrix of the sample eigenvector matrix is the largest, wherein the dimensionality reduction matrix is a 1 x 1 matrix, and the variance of the dimensionality reduction matrix is the product of sequential multiplication of the eigenvector weight matrix, the covariance matrix and a transposed matrix of the eigenvector weight matrix.
7. The apparatus according to claim 6, wherein the solving subunit solves the optimal solution of the eigenvector weight matrix under the condition that the preset constraint condition is satisfied by using a preset algorithm, specifically:
and solving the optimal solution of the eigenvector weight matrix under the condition of meeting the preset constraint condition by using a projection gradient descent method.
8. The apparatus according to claim 5, wherein the user behavior data obtaining unit obtains user behavior data of the target user for the target object, specifically:
acquiring a behavior log of the target user; determining user behavior data of the target user for the target object based on the behavior log.
9. An electronic device comprising a processor and a memory;
the processor is used for calling and executing the program stored in the memory;
the memory is configured to store the program, the program at least to:
acquiring user behavior data of a target user for a target object, wherein the user behavior data comprises positive feedback behavior data and negative feedback behavior data, the positive feedback behavior data represents that the user is interested in the target object, and the negative feedback behavior data represents that the user is not interested in the target object;
generating a multi-dimensional feature vector for the target user for the target object based on the user behavior data, each dimension of the multi-dimensional feature vector corresponding to one behavior data of the user behavior data;
processing the multi-dimensional feature vector by using a predetermined feature vector weight matrix to obtain a predicted value of the interest degree of the target user to the target object;
each dimension of the multi-dimensional feature vector corresponds to a feature vector weight in a feature vector weight matrix, and in the process of predetermining the feature vector weight matrix, a preset constraint condition needs to be met, wherein the preset constraint condition comprises: the weight of the feature vector corresponding to the positive feedback behavior data is greater than or equal to 0; the eigenvector weight corresponding to the negative feedback behavior data is less than or equal to 0.
10. A storage medium having stored thereon computer-executable instructions which, when loaded and executed by a processor, carry out a method of user interest analysis as claimed in any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910894901.2A CN110598949B (en) | 2019-09-20 | 2019-09-20 | User interest degree analysis method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910894901.2A CN110598949B (en) | 2019-09-20 | 2019-09-20 | User interest degree analysis method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110598949A true CN110598949A (en) | 2019-12-20 |
| CN110598949B CN110598949B (en) | 2021-10-15 |
Family
ID=68861941
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910894901.2A Active CN110598949B (en) | 2019-09-20 | 2019-09-20 | User interest degree analysis method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110598949B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113065837A (en) * | 2019-12-31 | 2021-07-02 | 深信服科技股份有限公司 | Method, device, electronic equipment and storage medium for determining user group |
| CN113127750A (en) * | 2021-05-19 | 2021-07-16 | 同盾科技有限公司 | Information list generation method and device, storage medium and electronic equipment |
| CN114780862A (en) * | 2022-06-21 | 2022-07-22 | 达而观数据(成都)有限公司 | User interest vector extraction method, extraction model and computer system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870972A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Data recommendation method and data recommendation system |
| CN104216960A (en) * | 2014-08-21 | 2014-12-17 | 北京奇艺世纪科技有限公司 | Method and device for recommending video |
| CN106127594A (en) * | 2016-06-30 | 2016-11-16 | 武汉斗鱼网络科技有限公司 | User room interest-degree computational methods based on principal component analysis and system |
| CN109815381A (en) * | 2018-12-21 | 2019-05-28 | 平安科技（深圳）有限公司 | User's portrait construction method, system, computer equipment and storage medium |
-
2019
- 2019-09-20 CN CN201910894901.2A patent/CN110598949B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870972A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Data recommendation method and data recommendation system |
| CN104216960A (en) * | 2014-08-21 | 2014-12-17 | 北京奇艺世纪科技有限公司 | Method and device for recommending video |
| CN106127594A (en) * | 2016-06-30 | 2016-11-16 | 武汉斗鱼网络科技有限公司 | User room interest-degree computational methods based on principal component analysis and system |
| CN109815381A (en) * | 2018-12-21 | 2019-05-28 | 平安科技（深圳）有限公司 | User's portrait construction method, system, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 雪柳花名（原文出处：中科春哥）: "主成成分（PCA）原理详解", 《HTTP://WWW.360DOC11.NET/WXARTICLENEW/681335432.HTML》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113065837A (en) * | 2019-12-31 | 2021-07-02 | 深信服科技股份有限公司 | Method, device, electronic equipment and storage medium for determining user group |
| CN113127750A (en) * | 2021-05-19 | 2021-07-16 | 同盾科技有限公司 | Information list generation method and device, storage medium and electronic equipment |
| CN113127750B (en) * | 2021-05-19 | 2024-04-02 | 同盾科技有限公司 | Information list generation method and device, storage medium and electronic equipment |
| CN114780862A (en) * | 2022-06-21 | 2022-07-22 | 达而观数据(成都)有限公司 | User interest vector extraction method, extraction model and computer system |
| CN114780862B (en) * | 2022-06-21 | 2022-09-23 | 达而观数据(成都)有限公司 | User interest vector extraction method, extraction model and computer system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110598949B (en) | 2021-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10699321B2 (en) | Global vector recommendations based on implicit interaction and profile data | |
| CN110321422B (en) | Method for training model on line, pushing method, device and equipment | |
| CN110598949B (en) | User interest degree analysis method and device, electronic equipment and storage medium | |
| US10609433B2 (en) | Recommendation information pushing method, server, and storage medium | |
| CN110807207B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN106952166B (en) | User influence estimation method and device of social platform | |
| US11379454B2 (en) | Automated social message stream population | |
| US20150227964A1 (en) | Revenue Estimation through Ensemble Modeling | |
| US20220261591A1 (en) | Data processing method and apparatus | |
| CN111159563B (en) | Method, device, equipment and storage medium for determining user interest point information | |
| CN110363580B (en) | Information recommendation method and device, computer equipment and storage medium | |
| US9633081B1 (en) | Systems and methods for determining application installation likelihood based on user network characteristics | |
| US20150310529A1 (en) | Web-behavior-augmented recommendations | |
| CN115082209A (en) | Business data risk early warning method and device, computer equipment and storage medium | |
| US20140289334A1 (en) | System and method for recommending multimedia information | |
| Sheng et al. | Parameter estimation in uncertain differential equations based on the solution | |
| Mitric et al. | On a risk measure inspired from the ruin probability and the expected deficit at ruin | |
| EP3293646A1 (en) | Method and server for training a machine learning algorithm (mla) for content recommendation generation | |
| CN110020129B (en) | Click rate correction method, prediction method, device, computing equipment and storage medium | |
| CN108009168B (en) | User account identification method and device | |
| CN111159558B (en) | Recommendation list generation method and device and electronic equipment | |
| CN109299351B (en) | Content recommendation method and device, electronic equipment and computer readable medium | |
| CN113515701A (en) | Information recommendation method and device | |
| US10353971B2 (en) | Increasing content circulation in a social networking service | |
| CN111831890A (en) | User similarity generation method and device, storage medium and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |