CN111737319A - User cluster prediction method and device, computer equipment and storage medium - Google Patents
User cluster prediction method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111737319A CN111737319A CN202010586411.9A CN202010586411A CN111737319A CN 111737319 A CN111737319 A CN 111737319A CN 202010586411 A CN202010586411 A CN 202010586411A CN 111737319 A CN111737319 A CN 111737319A
- Authority
- CN
- China
- Prior art keywords
- user
- cluster
- prediction
- prediction model
- feature
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000007499 fusion processing Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims description 27
- 238000004590 computer program Methods 0.000 claims description 25
- 238000012216 screening Methods 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000010801 machine learning Methods 0.000 description 10
- 230000009286 beneficial effect Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000013102 re-test Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Abstract
The application relates to the technical field of intelligent decision making, and provides a prediction method and device of a user cluster, computer equipment and a storage medium. The method comprises the following steps: acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions; respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model; and performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster. In addition, the invention also relates to a block chain technology, and the target user cluster can be stored in the block chain node. By adopting the method, the user information characteristic codes of a plurality of user characteristic dimensions of each user are comprehensively considered, and the prediction is carried out through a plurality of user prediction models, so that the accuracy of the predicted user cluster is improved.
Description
Technical Field
The present application relates to the field of intelligent decision making technologies, and in particular, to a prediction method and apparatus for a user cluster, a computer device, and a storage medium.
Background
With the popularization of machine learning, more and more fields are applied to machine learning to perform effective analysis on data of corresponding fields, such as user cluster prediction.
However, in the current prediction method for a user cluster, generally, single-dimensional user information of a user, such as historical service operation information, is obtained and input into a machine learning model, so as to determine whether the user is a target user through the machine learning model; by analogy, user clusters meeting the conditions can be predicted through the machine learning model; however, whether the user is the target user is often determined by the influence of multiple factors, and the accuracy of the predicted user cluster is easily low because the user information of a single dimension of the user is analyzed only by one machine learning model.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a user cluster prediction method, apparatus, computer device, and storage medium capable of improving accuracy of a predicted user cluster.
A prediction method for a cluster of users, the method comprising:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
In one embodiment, the obtaining the feature codes of the users in the initial user cluster includes:
acquiring user information of a plurality of user characteristic dimensions of each user in the initial user cluster;
coding the user information of the plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user;
and splicing the user information feature codes of the plurality of user feature dimensions of each user to obtain the feature codes of each user.
In one embodiment, the encoding the user information features of the multiple user feature dimensions of each user into the corresponding user prediction models respectively to obtain the predicted user clusters output by each user prediction model includes:
inquiring the corresponding relation between preset user characteristic dimensions and a user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one;
respectively inputting the user information feature codes of the plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user on each user by each user prediction model;
and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model on each user.
In one embodiment, the obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model for each user includes:
extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user;
screening out users with the prediction probability larger than the preset probability from the users respectively, and correspondingly using the users as target users output by the user prediction models;
and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
In one embodiment, the pre-trained user prediction model is trained by:
acquiring a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user;
coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user;
respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user;
according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model;
carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets a convergence condition;
and if the user prediction models meet the convergence condition, correspondingly taking the user prediction models as the pre-trained user prediction models.
In one embodiment, after performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster, the method further includes:
acquiring credit scores of all target users in the target user cluster;
if the credit score is greater than or equal to a preset score, acquiring a resource type corresponding to the credit score;
pushing the resources corresponding to the resource types to corresponding target users;
if the credit score is smaller than the preset score, generating risk reminding information corresponding to the credit score;
and pushing the risk reminding information to a corresponding target user.
In one embodiment, after performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster, the method further includes:
and uploading the target user cluster to a block chain.
A prediction apparatus for a user cluster, the apparatus comprising:
the characteristic code acquisition module is used for acquiring the characteristic codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
the predicted user cluster obtaining module is used for inputting the user information feature codes of the user feature dimensions of each user into corresponding pre-trained user prediction models respectively to obtain predicted user clusters output by each user prediction model;
and the target user cluster acquisition module is used for performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
According to the prediction method, the prediction device, the computer equipment and the storage medium of the user cluster, the user information feature codes of a plurality of user feature dimensions of each user in the initial user cluster are obtained; secondly, inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models respectively to obtain predicted user clusters output by each user prediction model; finally, fusion processing is carried out on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster; the method and the device achieve the purpose of obtaining the target user cluster according to the user information feature codes of the multiple user feature dimensions of each user in the initial user cluster, comprehensively consider the user information feature codes of the multiple user feature dimensions of each user, and predict through the multiple user prediction models, are beneficial to improving the accuracy of the predicted user cluster, and avoid the defect that the accuracy of the predicted user cluster is low due to the fact that the user information of a single dimension of each user is analyzed through only one machine learning model.
Drawings
FIG. 1 is a diagram of an application environment for a prediction method for a cluster of users in one embodiment;
FIG. 2 is a flowchart illustrating a method for predicting a user cluster in one embodiment;
FIG. 3 is a flowchart illustrating a prediction method for a user cluster in another embodiment;
FIG. 4 is a block diagram of an embodiment of a prediction device for a user cluster;
FIG. 5 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The prediction method of the user cluster provided by the application can be applied to the application environment shown in fig. 1. Wherein the terminal 110 communicates with the server 120 through a network. The terminal 110 collects user information of a plurality of user characteristic dimensions of each user in the initial user cluster, and sends the user information of the plurality of user characteristic dimensions of each user to the server 120; the server 120 performs encoding processing on the user information of the plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model; and performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster. The terminal 110 may be, but not limited to, various personal computers, notebook computers, smart phones, and tablet computers, and the server 120 may be implemented by an independent server or a server cluster formed by a plurality of servers.
In an embodiment, as shown in fig. 2, a prediction method for a user cluster is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps:
step S201, acquiring feature codes of all users in an initial user cluster; the feature encoding includes user information feature encoding of a plurality of user feature dimensions.
The initial user cluster refers to a user cluster in which a target user (such as a core user) needs to be screened out, such as a historical user cluster; the user characteristic dimension refers to a dimension for describing user information, such as user basic information, recent operation behavior of a user, service information concerned by the user, service information transacted by the user, and the like; the user information feature coding refers to a low-dimensional feature vector which is subjected to compression coding and used for representing low-level semantics of user information, and can be obtained through pre-trained feature embedding network model learning.
Specifically, the server acquires user information of a plurality of user characteristic dimensions of each user in an initial user cluster, and codes the user information of the plurality of user characteristic dimensions of each user through a pre-trained characteristic embedding network model to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; therefore, the method is beneficial to subsequently and respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into the corresponding pre-trained user prediction models to obtain the predicted user clusters output by each user prediction model.
For example, a user selects an initial user cluster on a user prediction interface provided by a terminal, where the initial cluster includes user information of a plurality of user feature dimensions of each user; the terminal responds to the selection operation of the user, obtains the user information of a plurality of user characteristic dimensions of each user in the initial cluster, generates a user cluster prediction request according to the user information of the plurality of user characteristic dimensions of each user in the initial cluster, and sends the user cluster prediction request to a corresponding server; the server analyzes the user cluster prediction request to obtain user information of a plurality of user characteristic dimensions of each user in the initial cluster, and codes the user information of the plurality of user characteristic dimensions of each user in the initial cluster according to a preset coding instruction to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user in the initial cluster.
Step S202, respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain the predicted user clusters output by each user prediction model.
The user prediction model is a neural network model, such as a convolutional neural network model, a deep learning network model, and the like, for identifying whether a user is a target user (such as a key user). And different user characteristic dimensions and corresponding user prediction models are different. In an actual scene, the user prediction model comprises a user prediction model with partial static attributes, a user prediction model with partial generalization attributes and a user prediction model with partial dynamic attributes; the user prediction model with the partial static attributes can mine potential promotion spaces of more users, the user prediction model with the partial generalization attributes fully avoids the problem that the training samples are indirectly overfitted due to excessive feature dimensions, and the user prediction model with the partial dynamic attributes can more intuitively express a recent intention tendency of the user, so that a better recommendation strategy is found.
Specifically, the server acquires a pre-trained user prediction model corresponding to each user characteristic dimension, then respectively inputs user information characteristic codes of a plurality of user characteristic dimensions of each user into the corresponding pre-trained user prediction models, and performs a series of neural network processing, such as convolution processing, pooling processing, full-connection processing and the like, on the user information characteristic codes of the corresponding user characteristic dimensions of each user through each pre-trained user prediction model to obtain a prediction result of each user by each user prediction model; and respectively screening out clusters formed by users meeting the conditions from the prediction results of each user prediction model on each user, and correspondingly outputting the clusters as the prediction user clusters output by each user prediction model. Therefore, the user information feature codes of a plurality of user feature dimensions of each user are comprehensively considered, and prediction is carried out through a plurality of user prediction models, so that the accuracy of the user cluster predicted subsequently is improved.
For example, assume that there are 3 users in the initial user cluster, which are user a, user B, and user C, respectively, the user information feature codes of the multiple user feature dimensions corresponding to user a are a1, a2, and A3, respectively, the user information feature codes of the multiple user feature dimensions corresponding to user B are B1, B2, and B3, and the user information feature codes of the multiple user feature dimensions corresponding to user C are C1, C2, and C3, respectively; secondly, inputting the user information characteristic code A1 of the user A, the user information characteristic code B1 of the user B and the user information characteristic code C1 of the user C into a pre-trained user prediction model a respectively to obtain the prediction results of the user prediction model a on the user A, the user B and the user C, wherein if the user A and the user B meet the conditions, the predicted user cluster output by the user prediction model a comprises the user A and the user B; similarly, the user information feature code A2 of the user A, the user information feature code B2 of the user B and the user information feature code C2 of the user C are respectively input into a pre-trained user prediction model B to obtain the prediction results of the user prediction model B on the user A, the user B and the user C, and if the user B and the user C meet the conditions, the predicted user cluster output by the user prediction model B comprises the user B and the user C; and respectively inputting the user information characteristic code A3 of the user A, the user information characteristic code B3 of the user B and the user information characteristic code C3 of the user C into a pre-trained user prediction model C to obtain the prediction results of the user prediction model C on the user A, the user B and the user C, wherein if the user A and the user B meet the conditions, the predicted user cluster output by the user prediction model C comprises the user A and the user B.
Step S203, the predicted user clusters output by the user prediction models are subjected to fusion processing, and a target user cluster corresponding to the initial user cluster is obtained.
The target user cluster comprises a plurality of target users.
Specifically, the server performs fusion processing on predicted user clusters output by the user prediction models to count the occurrence probability of each user in the predicted user clusters, and screens out users with the occurrence probability larger than a preset occurrence probability from the predicted user clusters as target users; and constructing a target user cluster according to the target user, wherein the target user cluster is used as a target user cluster corresponding to the initial user cluster. Therefore, the purpose of obtaining the target user cluster corresponding to the initial user cluster according to the predicted user clusters output by the user prediction models is achieved, and the defect that the accuracy of the predicted user clusters is low easily caused by analyzing the user information of a single dimension of the user through only one machine learning model is avoided.
Further, after the predicted user clusters output by the user prediction models are subjected to fusion processing to obtain a target user cluster corresponding to the initial user cluster, the method further includes: and uploading the target user cluster to the block chain. Specifically, after obtaining the user cluster, the server may further store the target user cluster in a node of a block chain, so as to ensure the privacy and security of the target user cluster.
For example, the server generates a block of the target user cluster by using a block chain technology, and stores the block of the target user cluster into a node of a block chain, so as to store the target user cluster through the block chain, thereby preventing the target user cluster from being tampered, and ensuring the security of the obtained target user cluster.
In the prediction method of the user cluster, user information feature codes of a plurality of user feature dimensions of each user in an initial user cluster are obtained; secondly, inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models respectively to obtain predicted user clusters output by each user prediction model; finally, fusion processing is carried out on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster; the method and the device achieve the purpose of obtaining the target user cluster according to the user information feature codes of the multiple user feature dimensions of each user in the initial user cluster, comprehensively consider the user information feature codes of the multiple user feature dimensions of each user, and predict through the multiple user prediction models, are beneficial to improving the accuracy of the predicted user cluster, and avoid the defect that the accuracy of the predicted user cluster is low due to the fact that the user information of a single dimension of each user is analyzed through only one machine learning model.
In an embodiment, the step S201 of obtaining the feature codes of the users in the initial user cluster includes: acquiring user information of a plurality of user characteristic dimensions of each user in an initial user cluster; coding the user information of a plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; and splicing the user information feature codes of a plurality of user feature dimensions of each user to obtain the feature codes of each user.
For example, the server extracts the user information of the multiple user feature dimensions of each user in the initial user cluster from a local database in which the user information of the multiple user feature dimensions of the user is stored; respectively inputting the user information of a plurality of user characteristic dimensions of each user in the initial user cluster into a pre-trained characteristic embedded network model, and coding the user information of the plurality of user characteristic dimensions of each user through the pre-trained characteristic embedded network model to obtain the user information characteristic codes of the plurality of user characteristic dimensions of each user; and splicing and combining the user information feature codes of a plurality of user feature dimensions of each user according to a preset splicing and combining sequence to obtain the spliced and combined user information feature codes which are used as the feature codes of each user.
In this embodiment, by obtaining the feature codes of each user in the initial user cluster, it is beneficial to subsequently and respectively input the user information feature codes of a plurality of user feature dimensions of each user into the corresponding pre-trained user prediction model, so as to obtain the predicted user cluster output by each user prediction model.
In an embodiment, in step S202, the encoding the user information features of the multiple user feature dimensions of each user into the corresponding user prediction models respectively to obtain the predicted user cluster output by each user prediction model, includes: inquiring the corresponding relation between preset user characteristic dimensions and a user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one; respectively inputting user information feature codes of a plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user prediction model for each user; and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model to each user.
The preset corresponding relation between the user characteristic dimension and the user prediction model means that the user characteristic dimension and the user prediction model have a one-to-one corresponding relation.
For example, the server obtains a corresponding relationship between a preset user characteristic dimension and a user prediction model from a local database, and obtains the user prediction models corresponding to a plurality of user characteristic dimensions one to one according to the corresponding relationship between the preset user characteristic dimension and the user prediction model; respectively inputting user information feature codes of a plurality of user feature dimensions of a user into user prediction models which correspond to the user feature dimensions one by one to obtain prediction results of the user by each user prediction model; by analogy, the prediction result of each user prediction model for each user can be obtained; respectively screening target users meeting the conditions from the prediction results of the user prediction models for the users; and constructing a corresponding user cluster according to the target users meeting the conditions, and correspondingly outputting the user cluster as a prediction user cluster output by each user prediction model.
In this embodiment, the user information feature codes of a plurality of user feature dimensions of each user are comprehensively considered, and prediction is performed through a plurality of user prediction models, which is beneficial to improving the accuracy of a subsequently predicted user cluster.
In one embodiment, obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model for each user includes: extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user; respectively screening out users with the prediction probability higher than the preset probability from all the users, and correspondingly outputting the users as target users of all the user prediction models; and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
The prediction probability is used to measure whether the user is a target user, for example, to determine whether the user is a key user.
In this embodiment, the target user output by each user prediction model is determined according to the prediction probability of each user prediction model for each user, which is beneficial to improving the accuracy of the predicted user cluster output by each user prediction model.
In one embodiment, the pre-trained user prediction model is trained by: acquiring a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user; coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user; respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user; according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model; carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets the convergence condition; and if each user prediction model meets the convergence condition, corresponding each user prediction model as each pre-trained user prediction model.
The user prediction model meeting the convergence condition means that the training times of the user prediction model reach the preset training times, or the loss value of the user prediction model is smaller than the preset loss value.
For example, the server calculates the loss value of each user prediction model according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user in combination with a preset loss function; comparing the loss value of the user prediction model with a preset loss value, and if the loss value of the user prediction model is greater than or equal to the preset loss value, calculating the network parameter update gradient of the user prediction model according to the loss value of the user prediction model; updating the network parameters of the user prediction model according to the network parameter updating gradient of the user prediction model, and training the user prediction model after the network parameters are updated again until the loss value obtained according to the user prediction model is smaller than the preset loss value, and taking the user prediction model as a pre-trained user prediction model; with reference to this method, individual pre-trained user prediction models can be obtained.
In this embodiment, by repeatedly training each user prediction model, it is beneficial to improve the accuracy of the predicted user cluster output by the user prediction model, so that the accuracy of the subsequently obtained target user cluster is improved, and the accuracy of the predicted user cluster is further improved.
In an embodiment, in step S203, after performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster, the method further includes: acquiring credit scores of all target users in a target user cluster; if the credit score is greater than or equal to the preset score, acquiring a resource type corresponding to the credit score; and pushing the resources corresponding to the resource types to the corresponding target users.
The credit score is used for measuring the credit degree of the user, and the higher the credit score is, the higher the credit degree of the user is; the lower the credit score, the lower the user's level of credit. The resource refers to a product or a business, in particular to a financial product or a financial business in the financial field; the resource type is used for expressing a product type or a service type, and different resource types correspond to different product types or service types.
Specifically, the server queries and obtains credit scores of all target users in a target user cluster from a local database in which credit scores of a plurality of users are stored; and comparing the credit score of each target user with a preset score, if the credit score is greater than or equal to the preset score, acquiring a resource type corresponding to the credit score from the local database, and pushing the resource corresponding to the resource type to the corresponding target user terminal so as to display the resource corresponding to the resource type, such as financial products, financial services and the like, through the target user terminal.
In this embodiment, when the credit score of the target user is greater than or equal to the preset score, the resource of the resource type corresponding to the credit score is pushed to the target user, so that the purpose of accurate pushing is achieved, and the accuracy of resource pushing is further improved.
In another embodiment, after obtaining the credit scores of the respective target users in the target user cluster, the method further includes: if the credit score is smaller than the preset score, generating risk reminding information corresponding to the credit score; and pushing the risk reminding information to the corresponding target user.
Specifically, the server compares the credit score of each target user with a preset score, if the credit score is smaller than the preset score, a preset risk reminding information template is obtained, risk reminding information corresponding to the credit score is generated according to the preset risk reminding information template, and the risk reminding information is pushed to the corresponding target user terminal, so that the risk reminding information is displayed through the target user terminal, and the user is reminded of avoiding risks in time.
In an embodiment, as shown in fig. 3, another prediction method for a user cluster is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps:
step S301, user information of a plurality of user characteristic dimensions of each user in the initial user cluster is obtained.
Step S302, the user information of a plurality of user characteristic dimensions of each user is coded, and user information characteristic codes of a plurality of user characteristic dimensions of each user are obtained.
Step S303, splicing the user information feature codes of the plurality of user feature dimensions of each user to obtain the feature codes of each user.
Step S304, inquiring the corresponding relation between the preset user characteristic dimension and the user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one.
Step S305, respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one by one, and obtaining the prediction results of each user prediction model for each user.
Step S306, extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user.
Step S307, respectively screening out users with prediction probabilities larger than the preset probability from the users, and correspondingly outputting the users as target users of the prediction models of the users.
Step S308, a cluster formed by target users output by each user prediction model is obtained and correspondingly used as a prediction user cluster output by each user prediction model.
Step S309, performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster, and uploading the target user cluster to the block chain.
In this embodiment, the purpose of obtaining a target user cluster according to the user information feature codes of the multiple user feature dimensions of each user in the initial user cluster is achieved, the user information feature codes of the multiple user feature dimensions of each user are comprehensively considered, and prediction is performed through multiple user prediction models, so that the accuracy of the predicted user cluster is improved, and the defect that the accuracy of the predicted user cluster is low due to the fact that the user information of a single dimension of the user is analyzed through only one machine learning model is overcome.
In an embodiment, the present application further provides an application scenario, where the application scenario applies the prediction method for the user cluster described above. Specifically, the application of the prediction method for the user cluster in the application scenario is as follows:
(1) receiving the existing supportable data and distinguishing the field service of the existing basic data on the data equipment; and setting targets and benchmarking tasks which are in accordance with the data field aiming at data and data states of different sources.
(2) And selecting different user prediction models aiming at data and data states from different sources, such as a user prediction model with partial static attributes, a user prediction model with partial generalization attributes and a user prediction model with partial dynamic attributes.
(3) And training each user prediction model through the training data and the prediction data so as to ensure the comprehensiveness of the model effect.
(4) Training and updating detection are carried out on each user prediction model according to a preset business logic model algorithm and a target, and each trained user prediction model is obtained; and obtaining the service logic result of each data fragment area through the calculation of each user prediction model.
(5) Setting each model label according to the service logic result, such as an operation list; and completing the tasks of list integration and supplementary deletion under the intervention of rules and other model rules.
(6) And (4) carrying out retest on the existing test set by the simulation issuing strategy, and repeatedly adjusting the parameters and logics of all links to achieve the result of optimizing the overall target.
(7) When a new data sample enters, the distribution of the data is detected, and whether the need of updating the model parameters exists is judged.
(8) In the using process, according to a preset model and a rule strategy, a list strategy result is obtained for global data. The list policy described above is formed by summarizing the model results output by each model, and specifically may be implemented by integrating lists, supplementing and deleting business rules, and obtaining matching of label policies or matching of model policies for different lists.
(9) And returning the mode, the parameter and the weight adjustment of the modeling link, the model result link and the list integration link according to the operation feedback result of the client.
(10) Furthermore, considering that each link has obvious decoupling capacity, each module can be reused and promoted. The new service requirements can be met by adding and deleting modules. For example, when there is a new service requirement, the data layout blocks may have partial intersection, and the corresponding subsequent models, lists, and tags may be directly multiplexed and updated and iterated along with subsequent use. Moreover, when the business logic of the model changes, for example, a new policy is added, a corresponding data plate, or a model, a rule, and the like can be directly added, so that the use of the new policy is compatible.
According to the embodiment, a set of reasonable and simple algorithm model and a data storage use frame are designed, so that the business operation capacity of a bank is improved, and the experience of customers is optimized and improved. More users can be comprehensively covered in a vast list group; more operation and maintenance strategies can be adopted, and personalized user service experience is achieved through calculation of the algorithm. Meanwhile, a complete and meticulous logic framework can be reasonably adjusted, added and deleted according to results; the model can continuously meet the requirement of updating iteration according to the migration of a user sample and business through dynamic adjustment; the addition and deletion ensure that each logic module in the link can be multiplexed, and new service output can be performed more quickly.
It should be understood that although the various steps in the flow charts of fig. 2-3 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-3 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
In one embodiment, as shown in fig. 4, a prediction apparatus for a user cluster is provided, which includes: a feature code obtaining module 410, a predicted user cluster obtaining module 420, and a target user cluster obtaining module 430, wherein:
a feature code obtaining module 410, configured to obtain feature codes of users in the initial user cluster; the feature encoding includes user information feature encoding of a plurality of user feature dimensions.
The predicted user cluster obtaining module 420 is configured to input the user information feature codes of the multiple user feature dimensions of each user into the corresponding pre-trained user prediction models, respectively, to obtain the predicted user clusters output by each user prediction model.
And the target user cluster obtaining module 430 is configured to perform fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
In an embodiment, the feature code obtaining module 410 is further configured to obtain user information of a plurality of user feature dimensions of each user in the initial user cluster; coding the user information of a plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; and splicing the user information feature codes of a plurality of user feature dimensions of each user to obtain the feature codes of each user.
In an embodiment, the predicted user cluster obtaining module 420 is further configured to query a corresponding relationship between a preset user characteristic dimension and a user prediction model, so as to obtain a user prediction model corresponding to a plurality of user characteristic dimensions one to one; respectively inputting user information feature codes of a plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user prediction model for each user; and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model to each user.
In an embodiment, the predicted user cluster obtaining module 420 is further configured to extract a prediction probability of each user prediction model for each user from a prediction result of each user prediction model for each user; respectively screening out users with the prediction probability higher than the preset probability from all the users, and correspondingly outputting the users as target users of all the user prediction models; and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
In one embodiment, the prediction apparatus of the user cluster further includes a model training module, configured to obtain a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user; coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user; respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user; according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model; carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets the convergence condition; and if each user prediction model meets the convergence condition, corresponding each user prediction model as each pre-trained user prediction model.
In one embodiment, the prediction apparatus of the user cluster further includes a resource pushing module, configured to obtain credit scores of each target user in the target user cluster; if the credit score is greater than or equal to the preset score, acquiring a resource type corresponding to the credit score; and pushing the resources corresponding to the resource types to the corresponding target users.
In one embodiment, the prediction apparatus of the user cluster further includes an information pushing module, configured to generate risk reminding information corresponding to the credit score if the credit score is smaller than a preset score; and pushing the risk reminding information to the corresponding target user.
In one embodiment, the prediction apparatus of the user cluster further includes an upload module, configured to upload the target user cluster into the block chain.
For specific limitations of the prediction apparatus of the user cluster, reference may be made to the above limitations of the prediction method of the user cluster, which are not described herein again. The modules in the prediction device of the user cluster can be wholly or partially implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 5. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing the feature codes of all users, target user clusters and other data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a prediction method for a cluster of users.
Those skilled in the art will appreciate that the architecture shown in fig. 5 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster.
In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring user information of a plurality of user characteristic dimensions of each user in an initial user cluster; coding the user information of a plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; and splicing the user information feature codes of a plurality of user feature dimensions of each user to obtain the feature codes of each user.
In one embodiment, the processor, when executing the computer program, further performs the steps of: inquiring the corresponding relation between preset user characteristic dimensions and a user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one; respectively inputting user information feature codes of a plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user prediction model for each user; and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model to each user.
In one embodiment, the processor, when executing the computer program, further performs the steps of: extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user; respectively screening out users with the prediction probability higher than the preset probability from all the users, and correspondingly outputting the users as target users of all the user prediction models; and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user; coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user; respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user; according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model; carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets the convergence condition; and if each user prediction model meets the convergence condition, corresponding each user prediction model as each pre-trained user prediction model.
In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring credit scores of all target users in a target user cluster; if the credit score is greater than or equal to the preset score, acquiring a resource type corresponding to the credit score; pushing the resources corresponding to the resource types to corresponding target users; if the credit score is smaller than the preset score, generating risk reminding information corresponding to the credit score; and pushing the risk reminding information to the corresponding target user.
In one embodiment, the processor, when executing the computer program, further performs the steps of: and uploading the target user cluster to a block chain.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring user information of a plurality of user characteristic dimensions of each user in an initial user cluster; coding the user information of a plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user; and splicing the user information feature codes of a plurality of user feature dimensions of each user to obtain the feature codes of each user.
In one embodiment, the computer program when executed by the processor further performs the steps of: inquiring the corresponding relation between preset user characteristic dimensions and a user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one; respectively inputting user information feature codes of a plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user prediction model for each user; and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model to each user.
In one embodiment, the computer program when executed by the processor further performs the steps of: extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user; respectively screening out users with the prediction probability higher than the preset probability from all the users, and correspondingly outputting the users as target users of all the user prediction models; and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user; coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user; respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user; according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model; carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets the convergence condition; and if each user prediction model meets the convergence condition, corresponding each user prediction model as each pre-trained user prediction model.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring credit scores of all target users in a target user cluster; if the credit score is greater than or equal to the preset score, acquiring a resource type corresponding to the credit score; pushing the resources corresponding to the resource types to corresponding target users; if the credit score is smaller than the preset score, generating risk reminding information corresponding to the credit score; and pushing the risk reminding information to the corresponding target user.
In one embodiment, the computer program when executed by the processor further performs the steps of: and uploading the target user cluster to the block chain.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A prediction method for a user cluster, the method comprising:
acquiring feature codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
respectively inputting the user information feature codes of a plurality of user feature dimensions of each user into corresponding pre-trained user prediction models to obtain predicted user clusters output by each user prediction model;
and performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
2. The method of claim 1, wherein the obtaining the feature codes of the users in the initial user cluster comprises:
acquiring user information of a plurality of user characteristic dimensions of each user in the initial user cluster;
coding the user information of the plurality of user characteristic dimensions of each user to obtain user information characteristic codes of the plurality of user characteristic dimensions of each user;
and splicing the user information feature codes of the plurality of user feature dimensions of each user to obtain the feature codes of each user.
3. The method according to claim 1, wherein the step of inputting the user information feature codes of the plurality of user feature dimensions of each user into the corresponding user prediction models respectively to obtain the predicted user clusters output by each user prediction model comprises:
inquiring the corresponding relation between preset user characteristic dimensions and a user prediction model to obtain the user prediction models corresponding to the plurality of user characteristic dimensions one to one;
respectively inputting the user information feature codes of the plurality of user feature dimensions of each user into user prediction models corresponding to the plurality of user feature dimensions one to obtain prediction results of each user on each user by each user prediction model;
and obtaining the predicted user cluster output by each user prediction model according to the prediction result of each user prediction model on each user.
4. The method according to claim 3, wherein obtaining the predicted user cluster output by each of the user prediction models according to the prediction result of each of the user prediction models for each of the users comprises:
extracting the prediction probability of each user prediction model to each user from the prediction result of each user prediction model to each user;
screening out users with the prediction probability larger than the preset probability from the users respectively, and correspondingly using the users as target users output by the user prediction models;
and acquiring a cluster formed by target users output by each user prediction model, and correspondingly using the cluster as a prediction user cluster output by each user prediction model.
5. The method of claim 1, wherein the pre-trained user prediction model is trained by:
acquiring a sample user training set; the sample user training set comprises user information of each characteristic dimension of a sample user and actual probability of the sample user;
coding the user information of each characteristic dimension of the sample user to obtain the user information characteristic code of each characteristic dimension of the sample user;
respectively inputting the user information feature codes of each feature dimension of the sample user into each corresponding user prediction model to obtain the prediction probability of each user prediction model to the sample user;
according to the prediction probability of each user prediction model to the sample user and the actual probability of the sample user, calculating the loss value of each user prediction model;
carrying out reverse training on each user prediction model according to the loss value of each user prediction model until each user prediction model meets a convergence condition;
and if the user prediction models meet the convergence condition, correspondingly taking the user prediction models as the pre-trained user prediction models.
6. The method according to any one of claims 1 to 5, wherein after performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster, the method further comprises:
acquiring credit scores of all target users in the target user cluster;
if the credit score is greater than or equal to a preset score, acquiring a resource type corresponding to the credit score;
pushing the resources corresponding to the resource types to corresponding target users;
if the credit score is smaller than the preset score, generating risk reminding information corresponding to the credit score;
and pushing the risk reminding information to a corresponding target user.
7. The method according to any one of claims 1 to 5, wherein after performing fusion processing on the predicted user clusters output by each user prediction model to obtain a target user cluster corresponding to the initial user cluster, the method further comprises:
and uploading the target user cluster to a block chain.
8. An apparatus for predicting a user cluster, the apparatus comprising:
the characteristic code acquisition module is used for acquiring the characteristic codes of all users in the initial user cluster; the feature codes comprise user information feature codes of a plurality of user feature dimensions;
the predicted user cluster obtaining module is used for inputting the user information feature codes of the user feature dimensions of each user into corresponding pre-trained user prediction models respectively to obtain predicted user clusters output by each user prediction model;
and the target user cluster acquisition module is used for performing fusion processing on the predicted user clusters output by the user prediction models to obtain a target user cluster corresponding to the initial user cluster.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010586411.9A CN111737319B (en) | 2020-06-24 | 2020-06-24 | User cluster prediction method, device, computer equipment and storage medium |
| PCT/CN2020/134579 WO2021115269A1 (en) | 2020-06-24 | 2020-12-08 | User cluster prediction method, apparatus, computer device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010586411.9A CN111737319B (en) | 2020-06-24 | 2020-06-24 | User cluster prediction method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111737319A true CN111737319A (en) | 2020-10-02 |
| CN111737319B CN111737319B (en) | 2024-03-01 |
Family
ID=72650892
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010586411.9A Active CN111737319B (en) | 2020-06-24 | 2020-06-24 | User cluster prediction method, device, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111737319B (en) |
| WO (1) | WO2021115269A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021115269A1 (en) * | 2020-06-24 | 2021-06-17 | 平安科技（深圳）有限公司 | User cluster prediction method, apparatus, computer device, and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107292528A (en) * | 2017-06-30 | 2017-10-24 | 阿里巴巴集团控股有限公司 | Vehicle insurance Risk Forecast Method, device and server |
| CN109948633A (en) * | 2017-12-20 | 2019-06-28 | 广东欧珀移动通信有限公司 | User gender prediction method, apparatus, storage medium and electronic equipment |
| CN111259256A (en) * | 2020-02-13 | 2020-06-09 | 腾讯科技（深圳）有限公司 | Content processing method, content processing device, computer-readable storage medium and computer equipment |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105868847A (en) * | 2016-03-24 | 2016-08-17 | 车智互联（北京）科技有限公司 | Shopping behavior prediction method and device |
| US11144825B2 (en) * | 2016-12-01 | 2021-10-12 | University Of Southern California | Interpretable deep learning framework for mining and predictive modeling of health care data |
| US10540446B2 (en) * | 2018-01-31 | 2020-01-21 | Jungle Disk, L.L.C. | Natural language generation using pinned text and multiple discriminators |
| CN110348581B (en) * | 2019-06-19 | 2023-08-18 | 平安科技（深圳）有限公司 | User feature optimizing method, device, medium and electronic equipment in user feature group |
| CN111737319B (en) * | 2020-06-24 | 2024-03-01 | 平安科技（深圳）有限公司 | User cluster prediction method, device, computer equipment and storage medium |
-
2020
- 2020-06-24 CN CN202010586411.9A patent/CN111737319B/en active Active
- 2020-12-08 WO PCT/CN2020/134579 patent/WO2021115269A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107292528A (en) * | 2017-06-30 | 2017-10-24 | 阿里巴巴集团控股有限公司 | Vehicle insurance Risk Forecast Method, device and server |
| CN109948633A (en) * | 2017-12-20 | 2019-06-28 | 广东欧珀移动通信有限公司 | User gender prediction method, apparatus, storage medium and electronic equipment |
| CN111259256A (en) * | 2020-02-13 | 2020-06-09 | 腾讯科技（深圳）有限公司 | Content processing method, content processing device, computer-readable storage medium and computer equipment |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021115269A1 (en) * | 2020-06-24 | 2021-06-17 | 平安科技（深圳）有限公司 | User cluster prediction method, apparatus, computer device, and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111737319B (en) | 2024-03-01 |
| WO2021115269A1 (en) | 2021-06-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20180240041A1 (en) | Distributed hyperparameter tuning system for machine learning | |
| CN112633962B (en) | Service recommendation method and device, computer equipment and storage medium | |
| CN110929879A (en) | Business decision logic updating method based on decision engine and model platform | |
| KR102359090B1 (en) | Method and System for Real-time Abnormal Insider Event Detection on Enterprise Resource Planning System | |
| CN112328909B (en) | Information recommendation method and device, computer equipment and medium | |
| CN113822315A (en) | Attribute graph processing method and device, electronic equipment and readable storage medium | |
| CN112905340A (en) | System resource allocation method, device and equipment | |
| CN113592605A (en) | Product recommendation method, device, equipment and storage medium based on similar products | |
| CN115936895A (en) | Risk assessment method, device and equipment based on artificial intelligence and storage medium | |
| CN113011895B (en) | Associated account sample screening method, device and equipment and computer storage medium | |
| CN111737319B (en) | User cluster prediction method, device, computer equipment and storage medium | |
| CN113569162A (en) | Data processing method, device, equipment and storage medium | |
| CN115630221A (en) | Terminal application interface display data processing method and device and computer equipment | |
| CN111651652B (en) | Emotion tendency identification method, device, equipment and medium based on artificial intelligence | |
| CN114372867A (en) | User credit verification and evaluation method and device and computer equipment | |
| CN114170000A (en) | Credit card user risk category identification method, device, computer equipment and medium | |
| CN115455276A (en) | Method and device for recommending object, computer equipment and storage medium | |
| CN114154048A (en) | Method and device for constructing recommendation model, electronic equipment and storage medium | |
| CN112905987B (en) | Account identification method, device, server and storage medium | |
| CN115630973A (en) | User data processing method, device, computer equipment and storage medium | |
| CN116933049A (en) | Feature selection method, device, electronic equipment and storage medium | |
| CN114238726A (en) | User classification method, device, equipment and storage medium | |
| CN116151956A (en) | Credit approval processing method and device | |
| CN117422534A (en) | Method and device for creating financial wind control model, storage medium and electronic equipment | |
| CN117113222A (en) | Data analysis model generation method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40031946 Country of ref document: HK |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |