Artificial Intelligence-Based Data Processing Method, Device, Medium, and Equipment

Abstract

The present application provides an artificial intelligence-based data processing method, an artificial intelligence-based data processing device, a computer-readable storage medium, and an electronic device; relating to the field of artificial intelligence technology; including: predicting the probability set of the age group to which each piece of user data of a target user corresponds based on a first prediction model; aggregating each piece of user data into each time period of a time series according to the time stamp of each piece of user data; determining the probability set of the age group of each time period; calculating the relevant parameter set of each time period according to the probability set of the age group of each time period; generating user characteristics according to the probability set of the age group of each time period and the relevant parameter set of each time period, and determining the determination basis corresponding to the user characteristics based on a second prediction model, and performing specific user identification on the target user according to the determination basis. It can be seen that by implementing the embodiments of the present application, the accuracy of specific user determination can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology. Specifically, it relates to a data processing method based on artificial intelligence, a data processing device based on artificial intelligence, a computer-readable storage medium, and an electronic device. Background Art

[0002] With the continuous development of terminal games, the playability of various types of games is getting higher and higher. However, the self-control of some specific users (such as teenagers) is generally weak. During the game-playing process, they are prone to being overly addicted to the game, which will cause varying degrees of damage to the body and mind of specific users. Therefore, more and more game providers add anti-addiction mechanisms to games to prevent specific users from being overly addicted to the game by means such as forced game exit. Generally, the background will obtain user behavior data and predict the probability that the current user is a specific user based on the user behavior data. If the probability is relatively high, the anti-addiction mechanism will be activated for them. However, the general prediction method is only based on multiple disordered user data, which is not accurate and does not consider the external factors affecting user data. Therefore, there may be a problem of low determination accuracy in determining specific users according to the above method.

[0003] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present application. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0004] The purpose of the present application is to provide a data processing method based on artificial intelligence, a data processing device based on artificial intelligence, a computer-readable storage medium, and an electronic device, which can represent each piece of user data of a target user in each time period of a time series, so as to realize multi-dimensional data analysis based on time series. Furthermore, according to the analysis result, the user characteristics of the target user can be determined as the basis for specific user determination, which can improve the accuracy of specific user determination.

[0005] Other features and advantages of the present application will become apparent through the following detailed description, or will be partially learned through the practice of the present application.

[0006] According to one aspect of the present application, there is provided a data processing method based on artificial intelligence, including:

[0007] Obtain each piece of user data of a target user and predict the probability set corresponding to each piece of user data belonging to different age groups based on a first prediction model;

[0008] Read the time stamps in each piece of user data and aggregate each piece of user data into each time period of a time series according to the time stamps of each piece of user data to obtain an aggregation result;

[0009] Determine the age group probability sets for each time period according to the aggregation result and the age group probability sets corresponding to each piece of user data respectively;

[0010] Calculate the relevant parameter sets for each time period according to the age group probability sets for each time period; wherein, the relevant parameter sets are used as multi-dimensional representations of the user data within the corresponding time periods;

[0011] Generate user characteristics based on the age group probability sets for each time period and the relevant parameter sets for each time period, and determine the judgment basis corresponding to the user characteristics based on the second prediction model. Identify the target user as a specific user according to the judgment basis. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0012] According to one aspect of the present application, there is provided an artificial intelligence-based data processing device, including: a parameter prediction unit, a parameter aggregation unit, a parameter determination unit, a parameter calculation unit, a feature generation unit, and a specific user determination unit, wherein:

[0013] The parameter prediction unit is configured to obtain each piece of user data of the target user and predict the age group probability sets corresponding to each piece of user data respectively based on the first prediction model;

[0014] The parameter aggregation unit is configured to read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result;

[0015] The parameter determination unit is configured to determine the age group probability sets for each time period according to the aggregation result and the age group probability sets corresponding to each piece of user data respectively;

[0016] The parameter calculation unit is configured to calculate the relevant parameter sets for each time period according to the age group probability sets for each time period; wherein, the relevant parameter sets are used as multi-dimensional representations of the user data within the corresponding time periods;

[0017] The feature generation unit is configured to generate user characteristics according to the age group probability sets for each time period and the relevant parameter sets for each time period;

[0018] The specific user determination unit is configured to determine the judgment basis corresponding to the user characteristics based on the second prediction model, and identify the target user as a specific user according to the judgment basis. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0019] In an exemplary embodiment of the present application, any one of the probability sets of the corresponding age groups for each piece of user data includes: probabilities used to represent that the corresponding user data belongs to the preset age groups. The parameter prediction unit obtains each piece of user data of the target user and predicts the probability sets of the corresponding age groups for each piece of user data based on the first prediction model, including:

[0020] Collect each piece of user data corresponding to the target user within a unit time;

[0021] Perform feature transformation on each piece of user data to obtain the feature vectors corresponding to each piece of user data;

[0022] Input the feature vectors corresponding to each piece of user data into the first probability prediction model in sequence, so that the first probability prediction model calculates the probability sets of the corresponding age groups for each piece of user data according to the feature vectors corresponding to each piece of user data.

[0023] In an exemplary embodiment of the present application, the parameter determination unit determines the probability sets of the corresponding age groups for each time period according to the aggregation result and the probability sets of the corresponding age groups for each piece of user data, including:

[0024] Determine the user data divided into each time period according to the aggregation result to obtain the data volume of each time period;

[0025] Determine the probability sets of the corresponding age groups for each time period according to the data volume of each time period and the probability sets of the corresponding age groups for each piece of user data.

[0026] In an exemplary embodiment of the present application, the parameter determination unit determines the probability sets of the corresponding age groups for each time period according to the data volume of each time period and the probability sets of the corresponding age groups for each piece of user data, including:

[0027] If there is a first type of time period with a data volume greater than 1 among each time period, then fuse the probability sets of the corresponding age groups of the user data within the first type of time period to obtain the probability set of the corresponding age group of the first type of time period;

[0028] If there is a second type of time period with a data volume equal to 1 among each time period, then determine the probability set of the corresponding age group of the user data within the second type of time period as the probability set of the corresponding age group of the second type of time period;

[0029] If there is a third type of time period with a data volume less than 1 among each time period, then determine the probability set of the corresponding age group of the third type of time period according to the probability set of the corresponding age group of the previous adjacent time period of the third type of time period.

[0030] In an exemplary embodiment of the present application, any one of the relevant parameter sets for each time period includes: a score and a ranking. The parameter calculation unit calculates the relevant parameter sets for each time period according to the probability sets of the age groups to which each time period belongs, including:

[0031] Calculate the score for each time period according to the probability sets of the age groups to which each time period belongs, and determine the ranking for each time period according to the scores of each time period, so as to obtain the relevant parameter sets for each time period.

[0032] In an exemplary embodiment of the present application, the parameter calculation unit determines the ranking for each time period according to the score of each time period, including:

[0033] Determine the scores of other users in each time period;

[0034] For the same time period, determine the ranking of the target user among other users according to the scores of other users and the score of the target user, and accordingly obtain the ranking of the target user in each time period.

[0035] In an exemplary embodiment of the present application, the parameter calculation unit calculates the score for each time period according to the probability sets of the age groups to which each time period belongs, including:

[0036] Perform in-set probability fusion on the probability sets of the age groups to which each time period belongs to obtain the reference probabilities corresponding to each time period;

[0037] Calculate the score for each time period according to the reference probabilities corresponding to each time period and the probability sets of the age groups to which they belong.

[0038] In an exemplary embodiment of the present application, the parameter calculation unit calculates the score for each time period according to the reference probabilities corresponding to each time period and the probability sets of the age groups to which they belong, including:

[0039] Calculate the probability mean and probability standard deviation for each time period according to the reference probabilities of each time period;

[0040] Calculate the score for each time period according to the reference probabilities, probability mean, and probability standard deviation of each time period.

[0041] In an exemplary embodiment of the present application, the feature generation unit generates user features according to the probability sets of the age groups to which each time period belongs and the relevant parameter sets for each time period, including:

[0042] Generate time parameters corresponding to each piece of user data according to the time stamp of each piece of user data; wherein, the time parameter is used to characterize whether the corresponding user data is generated on a working day;

[0043] Generate user features based on the time parameters corresponding to each piece of user data, the set of probability of belonging age groups for each time period, and the set of relevant parameters for each time period.

[0044] In an exemplary embodiment of the present application, the feature generation unit generates user features based on the time parameters corresponding to each piece of user data, the set of probability of belonging age groups for each time period, and the set of relevant parameters for each time period, including:

[0045] Determine the time parameters of each time period according to the time parameters corresponding to each piece of user data;

[0046] Generate user features based on the time parameters of each time period, the set of probability of belonging age groups for each time period, and the set of relevant parameters for each time period.

[0047] In an exemplary embodiment of the present application, the feature generation unit generates user features based on the time parameters of each time period, the set of probability of belonging age groups for each time period, and the set of relevant parameters for each time period, including:

[0048] Concatenate the rankings in the time parameters within each time period, the set of probability of belonging age groups for each time period, and the set of relevant parameters for each time period to obtain the sub-features corresponding to each time period;

[0049] Concatenate the sub-features corresponding to each time period in time series to obtain user features.

[0050] In an exemplary embodiment of the present application, the specific user determination unit determines the determination basis corresponding to the user features based on the second prediction model, and performs specific user identification on the target user according to the determination basis, including:

[0051] Input the user features into the second prediction model to calculate the probability of the specific user corresponding to the user features through the second prediction model as the determination basis;

[0052] If the probability of the specific user is within the preset threshold range, determine that the target user is a specific user.

[0053] According to one aspect of the present application, there is provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein, the processor is configured to execute the method of any one of the above by executing the executable instructions.

[0054] According to one aspect of the present application, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method of any one of the above is implemented.

[0055] According to an aspect of the present application, there is provided a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the various optional implementation manners described above.

[0056] The exemplary embodiments of the present application may have the following partial or all beneficial effects:

[0057] In the data processing method based on artificial intelligence provided by an exemplary embodiment of the present application, it is possible to obtain each piece of user data of the target user and predict the set of probabilities of the corresponding age groups for each piece of user data based on the first prediction model; read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result; determine the set of probabilities of the corresponding age groups for each time period according to the aggregation result and the set of probabilities of the corresponding age groups for each piece of user data; calculate the set of relevant parameters for each time period according to the set of probabilities of the corresponding age groups for each time period; generate user characteristics according to the set of probabilities of the corresponding age groups for each time period and the set of relevant parameters for each time period, and determine the determination basis corresponding to the user characteristics based on the second prediction model, and perform specific user identification on the target user according to the determination basis. According to the above description of the solution, on the one hand, the present application can represent each piece of user data of the target user in each time period of the time series, so as to realize multi-dimensional data analysis based on time series. Furthermore, according to the analysis result, the user characteristics of the target user can be determined as the basis for specific user determination, which can improve the accuracy of specific user determination. On the other hand, the present application can perform data analysis on the user data within each time period in units of each time period, so as to obtain the set of relevant parameters for each time period, so as to realize multi-dimensional evaluation of each time period, thereby improving the accuracy and depth of data analysis, so as to determine the user characteristics that can accurately describe the target user.

[0058] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. Description of the Drawings

[0059] The drawings here are incorporated into the specification and constitute a part of the specification, showing the embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1A schematic diagram of an exemplary system architecture of a data processing method based on artificial intelligence and a data processing device based on artificial intelligence to which the embodiments of the present application can be applied is shown;

[0061] Figure 2 A schematic diagram of the structure of a computer system of an electronic device suitable for implementing the embodiments of the present application is shown;

[0062] Figure 3 A schematic diagram of a specific user determination process in the prior art is shown schematically;

[0063] Figure 4 A flowchart of a data processing method based on artificial intelligence according to an embodiment of the present application is shown schematically;

[0064] Figure 5 Schematic diagrams of user characteristics corresponding to multiple users according to an embodiment of the present application are shown;

[0065] Figure 6 A schematic diagram of the module structure for implementing a data processing method based on artificial intelligence is shown;

[0066] Figure 7 A flowchart of a data processing method based on artificial intelligence according to an embodiment of the present application is shown schematically;

[0067] Figure 8 A schematic diagram of the module structure for implementing a data processing method based on artificial intelligence according to an embodiment of the present application is shown;

[0068] Figure 9 A flowchart of a data processing method based on artificial intelligence according to an embodiment of the present application is shown schematically;

[0069] Figure 10 A block diagram of the structure of a data processing device based on artificial intelligence according to an embodiment of the present application is shown schematically. Detailed implementation manners

[0070] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art. The features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of this application. However, those skilled in the art will recognize that one or more of the specific details may be omitted, or other methods, components, devices, steps, etc. may be used. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring the various aspects of this application.

[0071] In addition, the accompanying drawings are only schematic illustrations of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0072] Figure 1 A schematic diagram of a system architecture of an exemplary application environment of a data processing method and a data processing device based on artificial intelligence to which the embodiments of this application can be applied is shown.

[0073] As Figure 1 shown, the system architecture 100 may include one or more of the terminal devices 101, 102, 103, a network 104, and a server cluster 105. The network 104 serves as a medium for providing a communication link between the terminal devices 101, 102, 103 and the server cluster 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc. The terminal devices 101, 102, 103 may be various electronic devices with a display screen, including but not limited to desktop computers, portable computers, smartphones, and tablet computers, etc. It should be understood that Figure 1 the number of terminal devices, networks, and servers in

[0074] The data processing method based on artificial intelligence provided by the embodiments of the present application can be executed by any one of the terminal devices 101, 102, 103 or the servers in the server cluster 105. Correspondingly, the data processing device based on artificial intelligence is generally disposed in the servers of the server cluster 105 or the terminal devices 101, 102, 103. For example, in an exemplary embodiment, any one of the servers in the server cluster 105 can obtain each piece of user data of the target user and predict the set of probabilities of the corresponding age groups of each piece of user data based on the first prediction model; read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result; determine the set of probabilities of the corresponding age groups of each time period according to the aggregation result and the set of probabilities of the corresponding age groups of each piece of user data; calculate the set of relevant parameters of each time period according to the set of probabilities of the corresponding age groups of each time period; generate user characteristics according to the set of probabilities of the corresponding age groups of each time period and the set of relevant parameters of each time period, and determine the determination basis corresponding to the user characteristics based on the second prediction model, and perform specific user identification on the target user according to the determination basis.

[0075] Figure 2 FIG. shows a schematic structural diagram of a computer system of an electronic device suitable for implementing the embodiments of the present application.

[0076] It should be noted that Figure 2 The computer system 200 of the electronic device shown is only an example and should not bring any limitation to the functions and usage scope of the embodiments of the present application.

[0077] As Figure 2 shown, the computer system 200 includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 202 or the program loaded from the storage section 208 into the random access memory (RAM) 203. In the RAM 203, various programs and data required for system operation are also stored. The CPU 201, ROM 202, and RAM 203 are connected to each other through a bus 204. The input / output (I / O) interface 205 is also connected to the bus 204.

[0078] The following components are connected to the I / O interface 205: an input section 206 including a keyboard, a mouse, etc.; an output section 207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 208 including a hard disk, etc.; and a communication section 209 including a network interface card such as a LAN card, a modem, etc. The communication section 209 performs communication processing via a network such as the Internet. A driver 210 is also connected to the I / O interface 205 as needed. A removable medium 211 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is installed on the driver 210 as needed so that a computer program read therefrom can be installed into the storage section 208 as needed.

[0079] Specifically, according to an embodiment of the present application, the processes described below with reference to the flowchart can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes program codes for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication section 209, and / or installed from the removable medium 211. When the computer program is executed by a central processing unit (CPU) 201, various functions defined in the method and apparatus of the present application are executed.

[0080] Blockchain technology can be applied in the present application. Specifically, the determination result for the target user can be uploaded to the blockchain so that it can be called at any time when needed, which can make the determination result tamper-proof and ensure the absolute correctness of the determination result for the target user. Among them, the blockchain is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithm. The blockchain, essentially a decentralized database, is a string of data blocks generated by using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity (anti-counterfeiting) of the information and generate the next block. The blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer.

[0081] The underlying blockchain platform may include processing modules such as user management, basic services, smart contracts, and operation monitoring. Among them, the user management module is responsible for the identity information management of all blockchain participants, including maintaining the generation of public and private keys (account management), key management, and the maintenance of the correspondence between the real identity of the user and the blockchain address (permission management). And under authorization, it supervises and audits the transaction situations of certain real identities, and provides the rule configuration for risk control (risk control audit); the basic service module is deployed on all blockchain node devices, used to verify the validity of business requests, and records them in storage after consensus on valid requests. For a new business request, the basic service first performs interface adaptation parsing and authentication processing (interface adaptation), then encrypts the business information through a consensus algorithm (consensus management), transmits it to the shared ledger completely and consistently after encryption (network communication), and performs record storage; the smart contract module is responsible for the registration and issuance of contracts, as well as contract triggering and contract execution. Developers can define contract logic through a certain programming language, publish it to the blockchain (contract registration), trigger the execution by calling keys or other events according to the logic of the contract terms, complete the contract logic, and at the same time provide functions for contract upgrade and cancellation; the operation monitoring module is mainly responsible for the deployment, configuration modification, contract setting, cloud adaptation during the product release process, and the visual output of the real-time status during product operation, such as: alarm, monitoring network conditions, monitoring the health status of node devices, etc.

[0082] The platform product service layer provides the basic capabilities and implementation frameworks of typical applications. Developers can build on these basic capabilities and overlay the characteristics of the business to complete the blockchain implementation of the business logic. The application service layer provides application services based on the blockchain solution for business participants to use.

[0083] In addition, this application also applies artificial intelligence technology and machine learning algorithms, specifically manifested in the prediction of the probability set of the age group and the prediction of the judgment basis. Artificial Intelligence (AI) is to use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results of theory, methods, technologies, and application systems. In other words, artificial intelligence is a comprehensive technology in computer science. It attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. Artificial intelligence is also the study of the design principles and implementation methods of various intelligent machines, enabling the machines to have the functions of perception, reasoning, and decision-making.

[0084] Artificial intelligence technology is an interdisciplinary subject with a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0085] Machine Learning (ML) is an interdisciplinary subject that involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications cover all fields of artificial intelligence. Machine learning and deep learning usually include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal teaching learning.

[0086] The solution provided in the embodiments of this application relates to the machine learning algorithm of artificial intelligence, and is specifically illustrated through the following embodiments:

[0087] At present, as the age of intelligent terminal users gradually decreases and game applications become extremely rich, the anti-addiction mechanism has become the focus of attention of all parties. Especially for mobile game providers, accurately capturing teenage players among a large number of users can provide important reference for the anti-addiction mechanism.

[0088] In order to avoid the anti-addiction mechanism, teenage players sometimes actively adopt various means to circumvent or deceive various identity verification mechanisms of manufacturers, such as registering accounts using other adults' ID cards, borrowing their parents' mobile phones for gaming and consumption, etc. Such behaviors often make it difficult to ensure the coverage rate of explicit anti-addiction systems, and overly strict or complex verification mechanisms will also greatly affect the experience of adult users who account for a larger proportion.

[0089] To solve this problem, the machine learning algorithm of artificial intelligence is adopted in the prior art to achieve the prediction of teenagers based on user data. For specific details, please refer to Figure 3 , Figure 3 which schematically shows a schematic diagram of the specific user determination process in the prior art. As Figure 3 shown, the specific user determination process in the prior art can be achieved by executing a feature transformation module 310, a prediction determination model 320, a mean value calculation module 330, and a mean value threshold comparison module 340.

[0090] Specifically, each piece of user data of the target user collected can be input into the feature transformation module 310, so that the feature transformation module 310 calculates a feature vector for each piece of user data, and the feature vectors of each piece of user data are obtained. Furthermore, the feature vectors of each piece of user data can be input into the prediction and determination model 320, so that the prediction and determination model 320 predicts the probability of belonging to teenagers corresponding to each feature vector. Furthermore, the mean value calculation module 330 calculates the mean value of the probability of belonging to teenagers corresponding to each feature vector, and the mean value threshold comparison module 340 compares the mean value with the threshold. If the mean value is less than or equal to the threshold, it is determined that the target user does not belong to the specific user; if the mean value is greater than the threshold, it is determined that the target user belongs to the specific user.

[0091] It can be seen that Figure 3 In the manner shown, the similarities and differences between multiple pieces of user data are not taken into account. In various scenarios of collecting user behavior data, the behaviors of users at different times are likely to have no obvious consistency, and the prior art does not explicitly handle this phenomenon, but instead expects the feature transformation and prediction model to be able to obtain a sufficiently good probability prediction result. In addition, the above solution has a large demand for data and a long data collection time cycle. Therefore, it is easy to cause the online model to not be trained in a timely manner. In addition, the model screens users not entirely based on the pre-anchored criteria. Therefore, the results will be affected by the overall data of the same batch.

[0092] Based on the above problems, the present exemplary embodiment provides an artificial intelligence-based data processing method. Please refer to Figure 4 , Figure 4 which schematically shows a flowchart of an artificial intelligence-based data processing method according to an embodiment of the present application. As Figure 4 shown, the artificial intelligence-based data processing method may include: step S410 to step S450.

[0093] Step S410: Obtain each piece of user data of the target user and predict the probability set of the corresponding age group to which each piece of user data belongs based on the first prediction model.

[0094] Step S420: Read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result.

[0095] Step S430: Determine the probability set of the corresponding age group for each time period according to the aggregation result and the probability set of the corresponding age group to which each piece of user data belongs.

[0096] Step S440: Calculate a set of relevant parameters for each time period according to the probability set of the corresponding age group for each time period; wherein, the set of relevant parameters is used as a multi-dimensional representation of the user data within the corresponding time period.

[0097] Step S450: Generate user characteristics based on the age group probability sets of each time period and the relevant parameter sets of each time period, determine the judgment basis corresponding to the user characteristics based on the second prediction model, and identify the target user as a specific user according to the judgment basis. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0098] Implementation Figure 4 By implementing the method shown, each piece of user data of the target user can be represented in each time period of the time series, so that multi-dimensional data analysis based on time series can be realized. Furthermore, according to the analysis results, the user characteristics of the target user can be determined as the basis for specific user judgment, which can improve the accuracy of specific user judgment. In addition, the user data within each time period can be analyzed with each time period as a unit to obtain the relevant parameter sets of each time period, so as to realize the multi-dimensional evaluation of each time period, thereby improving the accuracy and depth of data analysis, so as to determine the user characteristics that can accurately describe the target user.

[0099] Next, the above steps of this exemplary embodiment will be described in more detail.

[0100] In step S410, each piece of user data of the target user is obtained and the age group probability sets respectively corresponding to the pieces of user data are predicted based on the first prediction model.

[0101] Specifically, the first prediction model can be a Long Short-Term Memory (LSTM) network model. LSTM is an artificial neural network model used to process sequential data. For the first prediction model, it can be trained with the labeled sample user data. The age group probability sets of the sample user data output by the first prediction model are compared with the labeling results, and the parameters in the first prediction model are optimized according to the comparison results.

[0102] Optionally, before obtaining each piece of user data of the target user and predicting the set of probabilities of the corresponding age groups for each piece of user data based on the first prediction model, the above method may further include: integrating the data acquisition module into the project program to perform data collection and reporting in a timely manner; furthermore, multiple pieces of user data of the target user within a unit time (e.g., 48 hours) can be collected from the user logs; preprocessing the multiple pieces of user data to obtain each piece of user data for inputting into the first prediction model. Among them, the preprocessing process is used to standardize the multiple pieces of user data of the target user and unify the representation form of the user data. The preprocessed user data is represented as serialized data in form. Furthermore, it may also include: screening out illegal data that does not meet the preset conditions from the multiple pieces of user data, such as data with unreliable information, data with illegal formats, and data with too little information content.

[0103] In addition, the target user can be any user to be determined. This application can be applied to the game anti-addiction system to identify specific users (such as underage users); among them, the game anti-addiction system is used to prevent minors from overplaying games, advocate healthy gaming habits, and help legal guardians understand whether the guardianship object participates in this online game and is protected by the anti-addiction system, etc. The operation mechanism of the game anti-addiction system can include: detecting whether the current player is a minor player, and if so, taking corresponding measures.

[0104] It should be noted that each user can correspond to multiple pieces of user data. Each piece of user data of each user is generated according to the user's operations. Each operation of the user during the game process will correspond to the generation of a piece of user data to record each operation of the user. For example, the user data can be generated based on voice signals, page jump logic, operation codes, etc., or directly obtained from built-in sensors. In addition, the multiple pieces of user data can correspond to a specific game scenario or any game scenario.

[0105] In addition, the set of probabilities of the corresponding age groups can include probabilities for multiple age groups. Each probability is used to represent the probability that the target user belongs to that age group. The sets of probabilities of the corresponding age groups for multiple pieces of user data corresponding to the same user can be different. Optionally, the set of probabilities of the corresponding age groups can include [probability of 8 - 14 years old, probability of 15 - 17 years old, probability of 18 - 19 years old, probability of 20 - 24 years old, probability of 25 - 29 years old, probability of 30 - 34 years old, probability of 35 - 39 years old, probability of >= 40 years old].

[0106] For the probability sets of the corresponding age groups for each piece of user data, for example, the pieces of user data corresponding to the target user (e.g., Zhang San) are user data A, user data B, and user data C respectively; among them, user data A, user data B, and user data C are generated based on user operation A, user operation B, and user operation C respectively.

[0107] Specifically, the probability set of the corresponding age group for user data A can be [the probability of 8 - 14 years old is 2%, the probability of 15 - 17 years old is 3%, the probability of 18 - 19 years old is 3%, the probability of 20 - 24 years old is 1%, the probability of 25 - 29 years old is 1%, the probability of 30 - 34 years old is 5%, the probability of 35 - 39 years old is 5%, >= 40 years old is 80%]. It can be seen that the user data A corresponding to the target user (e.g., Zhang San) can indicate that the probability of Zhang San being >= 40 years old is 80%.

[0108] The probability set of the corresponding age group for user data B can be [the probability of 8 - 14 years old is 3%, the probability of 15 - 17 years old is 2%, the probability of 18 - 19 years old is 1%, the probability of 20 - 24 years old is 1%, the probability of 25 - 29 years old is 3%, the probability of 30 - 34 years old is 5%, the probability of 35 - 39 years old is 5%, >= 40 years old is 80%]. It can be seen that the user data B corresponding to the target user (e.g., Zhang San) can also indicate that the probability of Zhang San being >= 40 years old is 80%.

[0109] The probability set of the corresponding age group for user data C can be [the probability of 8 - 14 years old is 90%, the probability of 15 - 17 years old is 4%, the probability of 18 - 19 years old is 1%, the probability of 20 - 24 years old is 1%, the probability of 25 - 29 years old is 1%, the probability of 30 - 34 years old is 1%, the probability of 35 - 39 years old is 1%, >= 40 years old is 1%]. It can be seen that the user data C corresponding to the target user (e.g., Zhang San) can indicate that the probability of Zhang San being 8 - 14 years old is 90%.

[0110] As an optional embodiment, any one of the probability sets of the corresponding age groups for each piece of user data includes: the probability used to characterize that the corresponding user data belongs to the preset age groups. Obtaining each piece of user data of the target user and predicting the probability sets of the corresponding age groups for each piece of user data based on the first prediction model includes: collecting each piece of user data corresponding to the target user within a unit time; performing feature transformation on each piece of user data to obtain the feature vectors corresponding to each piece of user data; and sequentially inputting the feature vectors corresponding to each piece of user data into the first probability prediction model, so that the first probability prediction model calculates the probability sets of the corresponding age groups for each piece of user data according to the feature vectors corresponding to each piece of user data.

[0111] Specifically, the probability set of the age group is an eight-dimensional vector, which can be expressed as {f1, …, f n , f i}, where i is the dimension of the age group, and f i is the probability that the corresponding user data belongs to the i-th age group, and the value of f i ranges from 0 to 1. The specific age groups are [8 - 14, 15 - 17, 18 - 19, 20 - 24, 25 - 29, 30 - 34, 35 - 39, >= 40]. It can be set that <= 14 years old is a teenager, and > 14 years old is a non-teenager.

[0112] For example, the feature vectors a1, …, a n obtained after feature transformation of each piece of user data can be input into the first probability prediction model in sequence, so that the first probability prediction model generates the probability set of the age group {f1, …, f n , f i} corresponding to each piece of user data respectively.

[0113] It can be seen that implementing this optional embodiment can calculate the probability set of the age group corresponding to each piece of user data according to the feature vector of each piece of user data, so as to improve the accuracy of determining that the predicted target user belongs to a specific user. In addition, by splitting the original prediction model into a feature transformation module and a prediction model, that is, the first probability prediction model only performs prediction, the update amount can be reduced, the time overhead of online model update can be reduced, and the model optimization efficiency can be improved.

[0114] In step S420, the timestamps in each piece of user data are read, and each piece of user data is aggregated into each time period of the time series according to the timestamps of each piece of user data, and an aggregation result is obtained.

[0115] Specifically, the time series is a preset sampling period (e.g., 30 days), and each time period in the time series is adjacent to each other, and adjacent time periods can be spliced to form a time series; wherein, the number of time periods in this application is not limited, and the length of the time period is also not limited. Preferably, the length is 12 hours. In addition, there may or may not be an overlap between adjacent time periods in the time series.

[0116] Furthermore, reading the timestamps in each piece of user data and aggregating each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result includes: dividing each piece of user data into the corresponding time period of the time series according to the timestamp of each piece of user data to achieve the aggregation of each piece of user data; wherein, the timestamps of each piece of user data are respectively used to represent the generation time of each piece of user data.

[0117] For example, if the time periods include 00:00 - 02:00, 02:00 - 04:00, 04:00 - 06:00, and each piece of user data includes user data with a timestamp of 01:05, user data with a timestamp of 01:30, and user data with a timestamp of 02:25, then the user data with timestamps of 01:05 and 01:30 can be classified into the time period 00:00 - 02:00, and the user data with a timestamp of 02:25 can be classified into the time period 02:00 - 04:00.

[0118] In step S430, according to the aggregation result and the probability sets of the age groups corresponding to each piece of user data, determine the probability sets of the age groups for each time period.

[0119] Specifically, the aggregation result may include: each time period containing different numbers (or the same number) of user data.

[0120] As an alternative embodiment, determining the probability sets of the age groups for each time period according to the aggregation result and the probability sets of the age groups corresponding to each piece of user data includes: determining the user data classified into each time period according to the aggregation result to obtain the data volume of each time period; determining the probability sets of the age groups for each time period according to the data volume of each time period and the probability sets of the age groups corresponding to each piece of user data.

[0121] For example, if the time periods include 00:00 - 02:00, 02:00 - 04:00, 04:00 - 06:00, according to the aggregation result, determining the user data classified into each time period, it can be determined that the data volume of user data belonging to 00:00 - 02:00 is 2, the data volume of user data belonging to 02:00 - 04:00 is 4, and the data volume of user data belonging to 04:00 - 06:00 is 3.

[0122] Furthermore, for determining the probability sets of the age groups for each time period according to the data volume of each time period and the probability sets of the age groups corresponding to each piece of user data, the calculation method for the probability set of the age group for each time period is the same. Therefore, taking the target time period in each time period as an example, the calculation method for the probability set of the age group of the target time period is described. The target time period can be any time period in each time period: average the probabilities of the same age group in the target time period to obtain the probability of belonging to each age group in the target time period, that is, the probability set of the age group of the target time period.

[0123] For example, if the target time period is 00:00 - 02:00, the amount of user data belonging to 00:00 - 02:00 is 2, and the probability set of the age group to which one user data belongs is [probability of 8 - 14 years old is 2%, probability of 15 - 17 years old is 3%, probability of 18 - 19 years old is 3%, probability of 20 - 24 years old is 1%, probability of 25 - 29 years old is 1%, probability of 30 - 34 years old is 5%, probability of 35 - 39 years old is 5%, probability of >= 40 years old is 80%]. The probability set of the age group to which another user data belongs is [probability of 8 - 14 years old is 3%, probability of 15 - 17 years old is 2%, probability of 18 - 19 years old is 1%, probability of 20 - 24 years old is 1%, probability of 25 - 29 years old is 3%, probability of 30 - 34 years old is 5%, probability of 35 - 39 years old is 5%, probability of >= 40 years old is 80%]. Furthermore, the probabilities for the same age group of 8 - 14 years old can be averaged, that is, (2% + 3%) / 2. In the same way, the probability of the age group to which the target time period belongs in each of [8 - 14, 15 - 17, 18 - 19, 20 - 24, 25 - 29, 30 - 34, 35 - 39, >= 40] can be calculated, that is, the probability set of the age group to which it belongs is [(2% + 3%) / 2, (3% + 2%) / 2, (3% + 1%) / 2, (1% + 1%) / 2, (1% + 3%) / 2, (5% + 5%) / 2, (5% + 5%) / 2, (80% + 80%) / 2].

[0124] Optionally, before determining the probability set of the age group to which each time period belongs based on the data volume of each time period and the probability set of the age group to which each piece of user data belongs, it may further include: if a disturbance item setting operation is detected, set the target age group corresponding to the disturbance item setting operation as the disturbance item; screen out the probability corresponding to the target age group in the probability set of the age group to which each piece of user data belongs, so as to improve the calculation accuracy of the probability set of the age group to which each time period belongs.

[0125] It can be seen that implementing this optional embodiment can achieve time modeling under the user dimension to explicitly encode and learn the behavior of users at different time periods, so as to better handle the problem of user switching between different time periods caused by the elongation of the time scale (for example, teenagers are playing in the previous hour and parents are playing in the next hour), thereby improving the accuracy of determining whether the target user belongs to a specific user.

[0126] As an optional embodiment, determining the probability set of the age group to which each time period belongs based on the data volume of each time period and the probability set of the age group to which each piece of user data belongs includes:

[0127] If there is a first type of time period in which the data volume is greater than 1, then fuse the age group probability sets of the user data within the first type of time period to obtain the age group probability set of the first type of time period;

[0128] If there is a second type of time period in which the data volume is equal to 1, then determine the age group probability set of the user data within the second type of time period as the age group probability set of the second type of time period;

[0129] If there is a third type of time period in which the data volume is less than 1, then determine the age group probability set of the third type of time period according to the age group probability set of the previous adjacent time period of the third type of time period.

[0130] Specifically, a first type of time period with a data volume greater than 1 is used to indicate that at least two pieces of user data are included within this time period. Similarly, a second type of time period with a data volume equal to 1 is used to indicate that only one piece of user data is included within this time period, and a third type of time period with a data volume less than 1 is used to indicate that no user data is included within this time period.

[0131] For a time period containing at least two pieces of user data, the age group probability sets of its user data can be averaged pairwise. That is, the sets are aligned vertically, the first elements of each set are averaged, the second elements of each set are averaged, and so on, to obtain the age group probability set of the time period containing at least two pieces of user data.

[0132] For a time period that does not contain any user data, the age group probability set of the previous adjacent time period / the next adjacent time period of this time period can be directly used as the age group probability set of the time period that does not contain any user data. This can avoid the situation of null values, reduce the impact of null values on the final prediction result, and improve the data smoothness between adjacent time periods.

[0133] Please refer to Figure 5 , Figure 5 which schematically shows the user feature diagrams corresponding to multiple users according to an embodiment of the present application. As Figure 5 shown, according to steps S410 to S450, the user features corresponding to user 1, user 2, and user 3 can be determined. When user 1, user 2, and user 3 are successively applied to steps S410 to S450, user 1, user 2, and user 3 can successively serve as the target users in steps S410 to S450. Therefore, the acquisition methods of the user features of user 1, user 2, and user 3 are the same.

[0134] Specifically, for user 1 / user 2 / user 3, the probability set of the age group corresponding to each piece of user data of user 1 / user 2 / user 3 can be predicted based on the first prediction model, which is represented as a vector f. The timestamp in each piece of user data is read and each piece of user data is aggregated into each time period of the time series according to the timestamp of each piece of user data to obtain the aggregation result, where the time direction of the time series is Figure 5 It is represented by a straight line with a pointing arrow, and the vertical division of the straight line with the pointing arrow by the dotted line can obtain different time periods in the time series, that is, time period 1, time period 2...time period n, where n is a positive integer.

[0135] If there are at least two pieces of user data, an aggregation operation is performed on the time period containing at least two pieces of user data, specifically including: the time period containing at least two pieces of user data can be averaged for the age group probability set. That is, the sets are aligned vertically, the first element of each set is averaged, the second element of each set is averaged, and so on, to obtain the age group probability set of the time period containing at least two pieces of user data.

[0136] If there is a situation that does not contain any user data, a filling operation is performed on the time period that does not contain any user data, specifically including: directly using the age group probability set of the previous adjacent time period / next adjacent time period of the time period as the age group probability set of the time period that does not contain any user data.

[0137] Furthermore, the probability sets of the age groups corresponding to time periods 1, 2, ..., n can be determined based on the aggregation results and the probability sets of the age groups corresponding to each piece of user data, and the scores s of each time period can be calculated based on the probability sets of the age groups corresponding to each time period. Based on the calculated scores s of user 1, user 2, and user 3 in each time period of the time series, a vertical comparison can be performed to obtain the ranking r of the scores. For example, the scores s (e.g., 3, 7, 6) corresponding to user 1, user 2, and user 3 in time period 1 can be compared to obtain the respective rankings r (e.g., 3, 1, 2) of user 1, user 2, and user 3 in time period 1. That is, user 2 with a score s of 7> user 3 with a score s of 6> user 1 with a score s of 3.

[0138] According to the above method, the scores s and rankings r of user 1, user 2 and user 3 in each time period can be determined.

[0139] In addition, time parameters t corresponding to each piece of user data can be generated according to the timestamps of each piece of user data of User 1 / User 2 / User 3; among them, the time parameter t is used to characterize whether the corresponding user data is generated on a working day. Furthermore, according to the time parameter t, the vector f, and the ranking r, parameter descriptions for each time period can be constructed. Furthermore, by sequentially splicing according to the time sequence of time period 1, time period 2... time period n, user features 1, 2, and 3 corresponding to User 1, User 2, and User 3 can be obtained respectively, and user features 1, 2, and 3 all have temporality.

[0140] It can be seen that implementing this optional embodiment can obtain the probability set of the age group to which each time period belongs through data aggregation and / or data filling, so that the user features that accurately represent the target user can be calculated, thereby improving the determination accuracy for specific users.

[0141] In step S440, a set of relevant parameters for each time period is calculated according to the probability set of the age group to which each time period belongs; among them, the set of relevant parameters is used as a multi-dimensional representation of the user data within the corresponding time period.

[0142] Specifically, the set of relevant parameters for each time period includes the score and ranking of the corresponding time period.

[0143] As an optional embodiment, any set of relevant parameters in the set of relevant parameters for each time period includes: score and ranking. Calculating the set of relevant parameters for each time period according to the probability set of the age group to which each time period belongs includes: calculating the score of each time period according to the probability set of the age group to which each time period belongs, and determining the ranking of each time period according to the score of each time period to obtain the set of relevant parameters for each time period.

[0144] Specifically, the score is used as a personalized representation of each time period, and the ranking is used to represent the ranking order of the score of the target user in this time period compared with the scores of other users in this time period.

[0145] It can be seen that implementing this optional embodiment can calculate the score and ranking corresponding to each time period, which is beneficial to representing the user data within this time period from multiple dimensions, so as to improve the temporality of user features and improve the accuracy of prediction results.

[0146] As an optional embodiment, calculating the score of each time period according to the probability set of the age group to which each time period belongs includes: performing probability fusion within the set of the probability set of the age group to which each time period belongs to obtain the reference probability corresponding to each time period; calculating the score of each time period according to the reference probability corresponding to each time period and the probability set of the age group to which each time period belongs.

[0147] Specifically, perform probability fusion within the probability sets of the respective age groups for each time period to obtain the reference probabilities corresponding to each time period, including: screening out the probabilities of redundant age groups (such as the probability of the age group 15 - 17) in the probability sets of the respective age groups according to the user-set operation, merging the age groups of the screening results, and performing probability fusion to achieve dimensionality reduction of the probability sets of the respective age groups, thereby obtaining the reference probabilities corresponding to each time period. Among them, the specific method of performing probability fusion on the screening results is: According to or Calculate the reference probability p corresponding to each time period.

[0148] It can be seen that by implementing this optional embodiment, through the internal probability fusion of the probability sets of the respective age groups for each time period, the probability sets of the respective age groups that were originally multi-dimensional can be reduced to two dimensions, thereby improving the calculation efficiency of the scoring.

[0149] As an optional embodiment, calculate the scores of each time period according to the reference probabilities corresponding to each time period and the probability sets of the respective age groups, including: calculating the probability mean and probability standard deviation of each time period according to the reference probabilities of each time period; calculating the scores of each time period according to the reference probabilities, probability mean, and probability standard deviation of each time period.

[0150] Specifically, calculating the probability mean and probability standard deviation of each time period according to the reference probabilities of each time period includes: calculating the probability mean and the probability standard deviation std{p} of each time period according to the reference probability p corresponding to each time period. Based on this, calculating the scores of each time period according to the reference probabilities, probability mean, and probability standard deviation of each time period includes: Based on the expression Calculate the scores of each time period wherein, is used to represent the score of the i-th user in the time period t, and p i is used to represent the reference probability p of the i-th user.

[0151] It can be seen that by implementing this optional embodiment, based on the above specific calculation method, the calculation accuracy of the scoring can be improved.

[0152] As an optional embodiment, determine the rankings of each time period according to the scores of each time period, including: determining the scores of other users in each time period; for the same time period, determining the ranking of the target user among other users according to the scores of other users and the score of the target user, and accordingly obtaining the rankings of the target user in each time period.

[0153] Specifically, for the same time period, the ranking of the target user among other users is determined based on the ratings of other users and the ratings of the target user, including: the ratings of all users for each time period t Sort them to obtain the ranking of the target user among other users for each time period t. Among them, the ranking of the target user among other users can ensure the unity of its horizontal scale by means such as fixing the number of users participating in the operation or normalization.

[0154] It can be seen that implementing this optional embodiment can refer to the user data of other users to determine the ranking of the target user in each time period, which is beneficial to improving the accuracy of the generated user characteristics and the determination accuracy of specific users.

[0155] In step S450, user characteristics are generated based on the probability set of the age groups to which each time period belongs and the parameter set related to each time period, and the determination basis corresponding to the user characteristics is determined based on the second prediction model. The target user is identified as a specific user according to the determination basis. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0156] Specifically, the specific user can be a user in a specific age group, such as a user aged 0 to 14.

[0157] In addition, the second prediction model can be a Long Short-Term Memory (LSTM) network model. LSTM is an artificial neural network model used to process sequence data. Optionally, the first prediction model and the second prediction model can also be neural network models such as Convolutional Neural Networks (CNN) or Recurrent Neural Networks (RNN). The embodiments of the present application are not limited thereto. In addition, the first prediction model and the second prediction model correspond to different network structures and different network parameters.

[0158] As an optional embodiment, generating user characteristics based on the probability set of the age groups to which each time period belongs and the parameter set related to each time period includes: generating time parameters corresponding to each piece of user data according to the time stamps of each piece of user data; wherein, the time parameters are used to characterize whether the corresponding user data is generated on a working day; generating user characteristics based on the time parameters corresponding to each piece of user data, the probability set of the age groups to which each time period belongs, and the parameter set related to each time period.

[0159] Specifically, the time stamps of each piece of user data are used to characterize the generation time of each piece of user data. Optionally, the time parameters are also used to characterize whether the corresponding user data is generated during the day or during game activity time;

[0160] Among them, time parameters corresponding to each piece of user data are generated according to the timestamps of each piece of user data, including: inputting each piece of user data into a cycle encoder, so that the cycle encoder reads the timestamps in each piece of user data and determines whether the date to which the user data belongs is a working day according to the timestamps; if so, setting the time parameter corresponding to the user data to 1, which can be expressed as t = 1; if not, setting the time parameter corresponding to the user data to 0, which can be expressed as t = 0. Optionally, the cycle encoder can convert user data into time parameters based on the one-hot encoding algorithm, where the one-hot encoding algorithm is a process of converting categorical variables into a form that is easy for machine learning algorithms to utilize.

[0161] It can be seen that implementing this optional embodiment can generate user features based on time parameters and a set of relevant parameters to more accurately represent the target user, thereby facilitating the improvement of the prediction accuracy of the determination probability for specific users.

[0162] As an optional embodiment, user features are generated according to the time parameters corresponding to each piece of user data, the probability set of the age group to which each time period belongs, and the set of relevant parameters for each time period, including: determining the time parameters for each time period according to the time parameters corresponding to each piece of user data; generating user features according to the time parameters for each time period, the probability set of the age group to which each time period belongs, and the set of relevant parameters for each time period.

[0163] Among them, determining the time parameters for each time period according to the time parameters corresponding to each piece of user data includes: if the time parameters corresponding to the user data belonging to the same time period are all the same value (e.g., 0 or 1), then determining the time parameter for that time period to be that value (e.g., 0 or 1).

[0164] Optionally, generating user features according to the time parameters for each time period, the probability set of the age group to which each time period belongs, and the set of relevant parameters for each time period includes: splicing at least one parameter (i.e., score and / or ranking) among the time parameter t within each time period, the probability set of the age group to which each time period belongs, and the set of relevant parameters to obtain sub-features corresponding to each time period; splicing the sub-features corresponding to each time period in time series to obtain user features. Among them, the probability set of the age group to which each time period belongs is an eight-dimensional vector, and a ten-dimensional sub-feature can be obtained through splicing with the set of relevant parameters of the time parameter.

[0165] It can be seen that implementing this optional embodiment can add time parameters to user features to characterize whether the time period is a working day or a non-working day, and adding the working day factor to user features can improve the determination accuracy of specific users.

[0166] As an alternative embodiment, user characteristics are generated based on the time parameters of each time period, the probability set of the age groups to which each time period belongs, and the relevant parameter set of each time period, including: concatenating the rankings in the time parameters within each time period, the probability set of the age groups to which each time period belongs, and the relevant parameter set of each time period to obtain sub-characteristics corresponding to each time period; concatenating the sub-characteristics corresponding to each time period in time series to obtain user characteristics.

[0167] Specifically, the user characteristics can be the concatenation result of N ten-dimensional vectors, where N is the number of sub-characteristics. Among them, concatenating the rankings in the time parameters within each time period, the probability set of the age groups to which each time period belongs, and the relevant parameter set of each time period to obtain sub-characteristics corresponding to each time period includes: concatenating the time parameter t within each time period, the probability set f of the age groups to which each time period belongs, and the ranking r in the relevant parameter set of each time period to obtain sub-characteristics (t - f - r) corresponding to each time period.

[0168] It can be seen that implementing this alternative embodiment can obtain user characteristics for characterizing the target user through serial concatenation of sub-characteristics, thereby facilitating the improvement of the determination accuracy of the target user and avoiding misjudging the target user who does not belong to a specific user as belonging to a specific user.

[0169] Please refer to Figure 6 , Figure 6 which schematically shows the module structure diagram for implementing the data processing method based on artificial intelligence. As Figure 6 shown, the module structure diagram may include: a feature transformation module 610, a period encoder 620, a first prediction model 630, a fusion module 640, a scoring module 650, a ranking module 660, a feature generation module 670, and a second prediction model 680. The module structure diagram can be used to obtain the specific user probability corresponding to the target user, and the specific user probability is used to represent the probability that the target user belongs to a specific user. For Figure 5 the shown User 1, User 2, and User 3, the module structure diagram can be used to calculate their respective corresponding specific user probabilities, that is, the target user can be any user such as User 1, User 2, or User 3.

[0170] Specifically, each piece of user data of the target user can be input into the feature transformation module 610, so that the feature transformation module 610 transforms each piece of user data of the target user into a feature vector of the target user and inputs it into the first prediction model 630. The first prediction model 630 can calculate the probability set of the age group to which each piece of user data belongs as the vector f. It is also possible to read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result. Furthermore, the probability set of the age group to which each piece of user data belongs can be input into the fusion module 640, so that the probability set of the age group to which each piece of user data belongs is fused into the binary probability set of each piece of user data, so that the scoring module 650 can calculate the reference probability corresponding to each time period according to the binary probability set of each piece of user data, and calculate the score of each time period according to the reference probability corresponding to each time period and the probability set of the age group to which it belongs. The ranking module 660 can determine the ranking of each time period according to the scores of each time period. In addition, the cycle encoder 620 can read the timestamps in each piece of user data and determine whether the date to which the user data belongs is a working day according to the timestamps. If so, the time parameter t corresponding to the user data is set to 1, which can be expressed as t = 1; if not, the time parameter t corresponding to the user data is set to 0. Furthermore, the rankings, time parameter t, and vector f of each time period (such as time period 1, time period 2... time period n, where n is a positive integer) can be concatenated to obtain the sub-features of each time period. The feature generation module 670 can concatenate the sub-features of each time period to obtain the user feature of the target user. The second prediction model 680 can predict the probability that the target user belongs to a specific user according to the user feature of the target user, that is, the specific user probability.

[0171] As an optional embodiment, based on the determination basis corresponding to the user feature determined by the second prediction model, the target user is identified as a specific user according to the determination basis, including: inputting the user feature into the second prediction model to calculate the specific user probability corresponding to the user feature through the second prediction model as the determination basis; if the specific user probability is within a preset threshold range (such as 80% - 100%), it is determined that the target user is a specific user.

[0172] Specifically, the specific user probability is used to indicate whether the target user is a specific user. If the specific user probability is not within the preset threshold range, it is determined that the target user is not a specific user. In the case where it is determined that the target user is a specific user, the target user can also be marked. Furthermore, further identity verification can be performed on the target user.

[0173] Based on this, after determining that the target user is a specific user, the above method may further include: outputting a prompt message for prompting that the target user has been identified as a specific user (e.g., the system determines that you are a teenage user); detecting whether the continuous play time of the target user exceeds a threshold, and if so, outputting a prompt message for prompting that the game will be forced to exit after a preset duration (e.g., you have played the game excessively, and the game will be forced to exit in 3 minutes).

[0174] In addition, after outputting a prompt message for prompting that the game will be forced to exit after a preset duration, the above method may further include: outputting a warning message, which is used to warn the user that if they cannot exit the game within the preset duration, a play report will be fed back to the target device (e.g., the parent device), and the play report may include a play time record.

[0175] Furthermore, after outputting a prompt message for prompting that the game will be forced to exit after a preset duration, the above method may further include: if it is detected that the target user's account has not exited the game after the preset duration, executing a forced exit instruction, increasing the forced play interval duration corresponding to the target user proportionally, and reducing the play duration corresponding to the target user proportionally.

[0176] For example, if the forced play interval duration is 4 hours and the play duration is 30 minutes, the target user can only play for 30 minutes every 4 hours. If a specific user cannot abide by the set play duration, the forced play interval duration can be increased to 5 hours and the play duration can be reduced to 20 minutes, so as to achieve anti-addiction control of the game for specific users and a warning effect on specific users.

[0177] Further, after outputting a prompt message for prompting that the target user has been identified as a specific user, the above method may further include: if an appeal request is detected, outputting an appeal information submission interface according to the appeal request; receiving the appeal information input by the user through the appeal information submission interface; re-evaluating the target user according to the appeal information and feedbacking the re-evaluation result; if the re-evaluation result can be used to indicate that the target user is a specific user, performing the above detection of whether the continuous play time of the target user exceeds the threshold; if the re-evaluation result is used to indicate that the target user is not a specific user, ending the process.

[0178] Even further, after re-evaluating the target user according to the appeal information, the above method may further include: optimizing the second prediction model according to the re-evaluation result.

[0179] It can be seen that implementing this optional embodiment can determine periodic user characteristics based on periodic user data, and predict the probability of specific users based on this, thereby improving the determination accuracy of specific users.

[0180] Please refer toFigure 7 , Figure 7 schematically shows a flowchart of an artificial intelligence-based data processing method according to an embodiment of the present application. As Figure 7 shown, it includes steps S710 to S760.

[0181] Step S710: Collect user data of the target user within a unit time. Specifically, the unit time can be a period, such as 30 days.

[0182] Step S720: Screen the user data and represent the screened user data as feature vectors and input them into the first prediction model. Specifically, screening the user data can obtain user data that meets the preset conditions.

[0183] Step S730: Based on the first prediction model, predict the probability set of the age group to which the user data of the target user belongs. Specifically, feature transformation can be performed on each piece of user data to obtain the feature vectors corresponding to each piece of user data; the feature vectors corresponding to each piece of user data are sequentially input into the first probability prediction model, so that the first probability prediction model calculates the probability set of the age group to which each piece of user data belongs according to the feature vectors corresponding to each piece of user data.

[0184] Step S740: Determine the probability set of the age group for each time period according to the probability set of the age group to which each piece of user data belongs, calculate the relevant parameter set for each time period according to the probability set of the age group for each time period, and generate the user characteristics of the target user according to the relevant parameter set for each time period. Specifically, the timestamp in each piece of user data can be read and each piece of user data is aggregated into each time period of the time series according to the timestamp of each piece of user data to obtain an aggregation result, and then the above-mentioned determination of the probability set of the age group for each time period is performed according to the aggregation result and the probability set of the age group to which each piece of user data belongs.

[0185] Step S750: Train the second prediction model according to the user characteristics of other users generated historically and the marking results of the user characteristics of other users.

[0186] Step S760: Determine whether the target user belongs to a specific user based on the user characteristics of the target user through the trained second prediction model. Specifically, the user characteristics can be input into the second prediction model to calculate the probability of the specific user corresponding to the user characteristics through the second prediction model as the determination basis; if the probability of the specific user is within the preset threshold range, it is determined that the target user is a specific user.

[0187] Please refer to Figure 8 , Figure 8Schematically shown is a schematic diagram of a module structure for implementing an artificial-intelligence-based data processing method according to an embodiment of the present application. As Figure 8 shown, the schematic diagram of the module structure may include: a data acquisition module 810, a data preprocessing module 820, a first probability prediction module 830, a feature aggregation module 840, a second probability prediction module 850, and a determination result determination module 860.

[0188] Specifically, the data acquisition module 810 is configured to acquire multiple pieces of user data of a target user within a unit time (e.g., 48 hours). The data preprocessing module 820 is configured to standardize the multiple pieces of user data of the target user to unify the representation form of the user data. The first probability prediction module 830 is configured to obtain each piece of user data of the target user and predict a probability set of the corresponding age group to which each piece of user data belongs based on a first prediction model. The feature aggregation module 840 is configured to read the timestamps in each piece of user data and aggregate each piece of user data into each time period of a time series according to the timestamps of each piece of user data to obtain an aggregation result, determine a probability set of the corresponding age group for each time period according to the aggregation result and the probability sets of the corresponding age groups to which each piece of user data belongs, and calculate a parameter set of each time period according to the probability sets of the corresponding age groups for each time period. The second probability prediction module 850 is configured to generate user features according to the probability sets of the corresponding age groups for each time period and the parameter set of each time period, and determine a determination basis corresponding to the user features based on a second prediction model. The determination result determination module 860 is configured to perform specific user identification on the target user according to the determination basis to clarify whether the target user belongs to a specific user.

[0189] Please refer to Figure 9 , Figure 9 Schematically shown is a flowchart of an artificial-intelligence-based data processing method according to an embodiment of the present application. As Figure 9 shown, the artificial-intelligence-based data processing method includes: step S900 to step S970.

[0190] Step S900: Acquire each piece of user data corresponding to a target user within a unit time, perform feature transformation on each piece of user data to obtain a feature vector corresponding to each piece of user data, and sequentially input the feature vectors corresponding to each piece of user data into a first probability prediction model, so that the first probability prediction model calculates a probability set of the corresponding age group to which each piece of user data belongs according to the feature vectors corresponding to each piece of user data.

[0191] Step S910: Read the timestamps in each piece of user data and aggregate each piece of user data into each time period of a time series according to the timestamps of each piece of user data to obtain an aggregation result.

[0192] Step S920: Determine the user data divided into each time period according to the aggregation result to obtain the data volume of each time period.

[0193] Step S930: If there is a type of time period with a data volume greater than 1 among all time periods, then fuse the age group probability sets to which the user data within the type of time period belongs to obtain the age group probability set of the type of time period; if there is a type of time period with a data volume equal to 1 among all time periods, then determine the age group probability set to which the user data within the type of time period belongs as the age group probability set of the type of time period; if there is a type of time period with a data volume less than 1 among all time periods, then determine the age group probability set of the type of time period according to the age group probability set of the previous adjacent time period of the type of time period.

[0194] Step S940: Perform in-set probability fusion on the age group probability sets of each time period to obtain the reference probabilities corresponding to each time period, and calculate the probability mean and probability standard deviation of each time period according to the reference probabilities of each time period, and then calculate the scores of each time period according to the reference probabilities, probability means and probability standard deviations of each time period.

[0195] Step S950: Determine the scores of other users in each time period. For the same time period, determine the ranking of the target user among other users according to the scores of other users and the scores of the target user, and thus obtain the ranking of the target user in each time period; among them, the scores and rankings both belong to the relevant parameter set, and different time periods correspond to different relevant parameter sets.

[0196] Step S960: Generate time parameters corresponding to each piece of user data according to the time stamps of each piece of user data, where the time parameters are used to represent whether the corresponding user data is generated on a working day. Furthermore, determine the time parameters of each time period according to the time parameters corresponding to each piece of user data, and then splice the time parameters within each time period, the age group probability sets of each time period, and the rankings in the relevant parameter sets of each time period to obtain the sub-features corresponding to each time period. Splice the sub-features corresponding to each time period in time series to obtain the user features.

[0197] Step S970: Input the user features into the second prediction model to calculate the specific user probability corresponding to the user features through the second prediction model as the judgment basis. If the specific user probability is within the preset threshold range, then determine that the target user is a specific user. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0198] It should be noted that Step S900 to Step S970 correspond to Figure 4 the steps and their embodiments shown. For the specific implementation manners of Step S900 to Step S970, please refer toFigure 4 The various steps and their embodiments shown are not elaborated here.

[0199] It can be seen that implementing Figure 9 the method shown can represent each piece of user data of the target user in each time period of the time series, so as to realize multi-dimensional data analysis based on temporality. Furthermore, based on the analysis results, the user characteristics of the target user can be determined as the basis for specific user determination, which can improve the accuracy of specific user determination. In addition, it is possible to perform data analysis on the user data within each time period in units of each time period, so as to obtain a set of relevant parameters for each time period to realize multi-dimensional evaluation of each time period, thereby improving the accuracy and depth of data analysis in order to determine the user characteristics that can accurately describe the target user.

[0200] Furthermore, in the present exemplary embodiment, a data processing device based on artificial intelligence is also provided. Referring to Figure 10 shown, the data processing device 1000 based on artificial intelligence may include: a parameter prediction unit 1001, a parameter aggregation unit 1002, a parameter determination unit 1003, a parameter calculation unit 1004, a feature generation unit 1005, and a specific user determination unit 1006, where:

[0201] The parameter prediction unit 1001 is configured to obtain each piece of user data of the target user and predict the probability set of the age groups to which the respective pieces of user data belong based on the first prediction model;

[0202] The parameter aggregation unit 1002 is configured to read the timestamps in each piece of user data and aggregate each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result;

[0203] The parameter determination unit 1003 is configured to determine the probability set of the age groups to which each time period belongs according to the aggregation result and the probability set of the age groups to which each piece of user data belongs;

[0204] The parameter calculation unit 1004 is configured to calculate a set of relevant parameters for each time period according to the probability set of the age groups to which each time period belongs; wherein, the set of relevant parameters is used as a multi-dimensional representation of the user data within the corresponding time period;

[0205] The feature generation unit 1005 is configured to generate user characteristics according to the probability set of the age groups to which each time period belongs and the set of relevant parameters for each time period;

[0206] A specific user determination unit 1006 is configured to determine a determination basis corresponding to user characteristics based on a second prediction model, and identify a target user as a specific user according to the determination basis. The first probability prediction model and the second probability prediction model correspond to different model parameters.

[0207] It can be seen that implementing Figure 10 the shown device can represent each piece of user data of the target user in each time period of the time series, so that multi-dimensional data analysis based on temporality can be achieved. Furthermore, according to the analysis results, the user characteristics of the target user can be determined as the determination basis for specific users, which can improve the accuracy of specific user determination. In addition, data analysis can be performed on the user data within each time period in units of each time period, so as to obtain a set of relevant parameters for each time period to achieve multi-dimensional evaluation of each time period, thereby improving the accuracy and depth of data analysis, so as to determine the user characteristics that can accurately describe the target user.

[0208] In an exemplary embodiment of the present application, any one of the probability sets of the corresponding age groups of each piece of user data includes: probabilities used to represent that the corresponding user data belongs to the preset age groups. The parameter prediction unit 1001 obtains each piece of user data of the target user and predicts the probability sets of the corresponding age groups of each piece of user data based on the first prediction model, including:

[0209] Collect each piece of user data corresponding to the target user within a unit time;

[0210] Perform feature transformation on each piece of user data to obtain feature vectors corresponding to each piece of user data;

[0211] Input the feature vectors corresponding to each piece of user data into the first probability prediction model in sequence, so that the first probability prediction model calculates the probability sets of the corresponding age groups of each piece of user data according to the feature vectors corresponding to each piece of user data.

[0212] It can be seen that implementing this optional embodiment can calculate the probability sets of the corresponding age groups of each piece of user data according to the feature vectors of each piece of user data, so as to improve the accuracy of determining and predicting that the target user belongs to a specific user. In addition, by splitting the original prediction model into a feature transformation module and a prediction model, that is, the first probability prediction model only performs prediction, the update amount can be reduced, the time overhead of online model update can be reduced, and the model optimization efficiency can be improved.

[0213] In an exemplary embodiment of the present application, the parameter determination unit 1003 determines the probability sets of the corresponding age groups of each time period according to the aggregation result and the probability sets of the corresponding age groups of each piece of user data, including:

[0214] Determine the user data divided into each time period according to the aggregation result to obtain the data volume of each time period;

[0215] Determine the probability set of the age group to which each time period belongs according to the data volume of each time period and the probability set of the age group to which each piece of user data corresponds.

[0216] It can be seen that implementing this optional embodiment can achieve time modeling under the user dimension to explicitly encode and learn the behaviors of users in different time periods, so as to better handle the problem of user switching between different time periods caused by the elongation of the time scale (for example, teenagers are playing in the previous hour and parents are playing in the next hour), thereby improving the accuracy of determining whether the target user belongs to a specific user.

[0217] In an exemplary embodiment of the present application, the parameter determination unit 1003 determines the probability set of the age group to which each time period belongs according to the data volume of each time period and the probability set of the age group to which each piece of user data corresponds, including:

[0218] If there is a type of time period in which the data volume is greater than 1 among each time period, fuse the probability sets of the age groups to which the user data in the type of time period belongs to obtain the probability set of the age group to which the type of time period belongs;

[0219] If there is a type of time period in which the data volume is equal to 1 among each time period, determine the probability set of the age group to which the user data in the type of time period belongs as the probability set of the age group to which the type of time period belongs;

[0220] If there is a type of time period in which the data volume is less than 1 among each time period, determine the probability set of the age group to which the type of time period belongs according to the probability set of the age group to which the previous adjacent time period of the type of time period belongs.

[0221] It can be seen that implementing this optional embodiment can obtain the probability set of the age group to which each time period belongs through data aggregation and / or data filling, so that the user characteristics accurately representing the target user can be calculated, thereby improving the determination accuracy for specific users.

[0222] In an exemplary embodiment of the present application, any relevant parameter set in the relevant parameter sets of each time period includes: score and ranking. The parameter calculation unit 1004 calculates the relevant parameter sets of each time period according to the probability set of the age group to which each time period belongs, including:

[0223] Calculate the score of each time period according to the probability set of the age group to which each time period belongs, and determine the ranking of each time period according to the score of each time period to obtain the relevant parameter sets of each time period.

[0224] It can be seen that implementing this optional embodiment can calculate the scores and rankings corresponding to each time period, thereby facilitating the characterization of user data within that time period from multiple dimensions to enhance the temporality of user features and improve the accuracy of prediction results.

[0225] In an exemplary embodiment of the present application, the parameter calculation unit 1004 determines the ranking of each time period according to the scores of each time period, including:

[0226] Determine the scores of other users in each time period;

[0227] For the same time period, determine the ranking of the target user among other users according to the scores of other users and the scores of the target user, and thus obtain the ranking of the target user in each time period.

[0228] It can be seen that implementing this optional embodiment can refer to the user data of other users to determine the ranking of the target user in each time period, thereby facilitating the improvement of the accuracy of the generated user features and the accuracy of specific user determination.

[0229] In an exemplary embodiment of the present application, the parameter calculation unit 1004 calculates the score of each time period according to the probability set of the age group to which each time period belongs, including:

[0230] Perform probability fusion within the probability set of the age group to which each time period belongs to obtain the reference probability corresponding to each time period;

[0231] Calculate the score of each time period according to the reference probability corresponding to each time period and the probability set of the age group to which it belongs.

[0232] It can be seen that implementing this optional embodiment can reduce the probability set of the age group to which each time period belongs, which is originally multi-dimensional, to two-dimensional through internal probability fusion of the probability set of the age group to which each time period belongs, thereby improving the calculation efficiency of the score.

[0233] In an exemplary embodiment of the present application, the parameter calculation unit 1004 calculates the score of each time period according to the reference probability corresponding to each time period and the probability set of the age group to which it belongs, including:

[0234] Calculate the probability mean and probability standard deviation of each time period according to the reference probability of each time period;

[0235] Calculate the score of each time period according to the reference probability, probability mean and probability standard deviation of each time period.

[0236] It can be seen that implementing this optional embodiment can improve the calculation accuracy of the score based on the above specific calculation method.

[0237] In an exemplary embodiment of the present application, the feature generation unit 1005 generates user features according to the probability set of the age group to which each time period belongs and the parameter set related to each time period, including:

[0238] Generating a time parameter corresponding to each piece of user data according to the timestamp of each piece of user data; wherein, the time parameter is used to characterize whether the corresponding user data is generated on a working day;

[0239] Generating user features according to the time parameter corresponding to each piece of user data, the probability set of the age group to which each time period belongs, and the parameter set related to each time period.

[0240] It can be seen that implementing this optional embodiment can generate user features based on the time parameter and the parameter set, so as to more accurately characterize the target user, thereby facilitating the improvement of the prediction accuracy of the determination probability for a specific user.

[0241] In an exemplary embodiment of the present application, the feature generation unit 1005 generates user features according to the time parameter corresponding to each piece of user data, the probability set of the age group to which each time period belongs, and the parameter set related to each time period, including:

[0242] Determining the time parameter of each time period according to the time parameter corresponding to each piece of user data;

[0243] Generating user features according to the time parameter of each time period, the probability set of the age group to which each time period belongs, and the parameter set related to each time period.

[0244] It can be seen that implementing this optional embodiment can add a time parameter to the user features to characterize whether the time period is a working day or a non - working day, and adding the working day factor to the user features can improve the determination accuracy of a specific user.

[0245] In an exemplary embodiment of the present application, the feature generation unit 1005 generates user features according to the time parameter of each time period, the probability set of the age group to which each time period belongs, and the parameter set related to each time period, including:

[0246] Concatenating the rankings in the time parameter within each time period, the probability set of the age group to which each time period belongs, and the parameter set related to each time period to obtain a sub - feature corresponding to each time period;

[0247] Concatenating the sub - features corresponding to each time period in time series to obtain user features.

[0248] It can be seen that by implementing this optional embodiment, user features characterizing the target user can be obtained through serial splicing of sub-features, which helps to improve the determination accuracy of the target user and avoid misjudging the target user who does not belong to a specific user as belonging to the specific user.

[0249] In an exemplary embodiment of the present application, the specific user determination unit 1006 determines the determination basis corresponding to the user features based on the second prediction model, and performs specific user identification on the target user according to the determination basis, including:

[0250] Input the user features into the second prediction model to calculate the probability of the target user being a specific user corresponding to the user features through the second prediction model as the determination basis;

[0251] If the probability of the target user being a specific user is within the preset threshold range, it is determined that the target user is a specific user.

[0252] It can be seen that by implementing this optional embodiment, periodic user features can be determined based on periodic user data, and the probability of a specific user can be predicted based on this, thereby improving the determination accuracy of a specific user.

[0253] It should be noted that although several modules or units of a device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of the two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0254] Since each functional module of the data processing device based on artificial intelligence in the exemplary embodiment of the present application corresponds to the steps of the exemplary embodiment of the above-mentioned artificial intelligence-based data processing method, for the details not disclosed in the device embodiment of the present application, please refer to the embodiments of the above-mentioned artificial intelligence-based data processing method of the present application.

[0255] On the other hand, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist alone without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the one or more programs are executed by an electronic device, the electronic device realizes the method described in the above embodiment.

[0256] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. And in this application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0257] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the above-mentioned module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, as well as the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0258] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not, in some cases, constitute a limitation on the unit itself.

[0259] Those skilled in the art will readily conceive of other embodiments of this application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application, which follow the general principles of this application and include known common general knowledge or conventional technical means in the art not disclosed in this application. The specification and examples are only to be considered exemplary, and the true scope and spirit of this application are pointed out by the following claims.

[0260] It should be understood that this application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is only limited by the appended claims.

Claims

1. A data processing method based on artificial intelligence, characterized in that, Including: Obtaining each piece of user data of a target user and predicting the set of probabilities of the corresponding age groups for each piece of user data based on a first prediction model; Reading the timestamps in each piece of user data and aggregating each piece of user data into each time period of a time series according to the timestamps of each piece of user data to obtain an aggregation result; Determining the set of probabilities of the corresponding age groups for each time period according to the aggregation result and the set of probabilities of the corresponding age groups for each piece of user data; Calculating the set of relevant parameters for each time period according to the set of probabilities of the corresponding age groups for each time period; wherein, any set of relevant parameters in the set of relevant parameters for each time period includes: a score and a ranking, the score is used as a personalized representation of each time period, and the ranking is used to represent the ranking order of the score of the target user in the time period compared with the scores of other users in the time period; Generating user features according to the set of probabilities of the corresponding age groups for each time period and the set of relevant parameters for each time period; Determining the determination basis corresponding to the user features based on a second prediction model, and performing specific user identification on the target user according to the determination basis, wherein the first prediction model and the second prediction model correspond to different model parameters.

2. The method according to claim 1, wherein Any set of probabilities of the corresponding age groups in the set of probabilities of the corresponding age groups for each piece of user data includes: the probability used to represent that the corresponding user data belongs to each preset age group. Obtaining each piece of user data of a target user and predicting the set of probabilities of the corresponding age groups for each piece of user data based on a first prediction model includes: Collecting each piece of user data corresponding to the target user within a unit time; Performing feature transformation on each piece of user data to obtain the feature vector corresponding to each piece of user data; Sequentially inputting the feature vectors corresponding to each piece of user data into the first prediction model, so that the first prediction model calculates the set of probabilities of the corresponding age groups for each piece of user data according to the feature vectors corresponding to each piece of user data.

3. The method according to claim 1, characterized in that Determining the set of probabilities of the corresponding age groups for each time period according to the aggregation result and the set of probabilities of the corresponding age groups for each piece of user data includes: Determining the user data divided into each time period according to the aggregation result to obtain the data volume of each time period; Determining the set of probabilities of the corresponding age groups for each time period according to the data volume of each time period and the set of probabilities of the corresponding age groups for each piece of user data.

4. The method according to claim 3, wherein Determining the set of probabilities of the corresponding age groups for each time period according to the data volume of each time period and the set of probabilities of the corresponding age groups for each piece of user data includes: If there is a type of time period with a data volume greater than 1 among each time period, then fusing the set of probabilities of the corresponding age groups of the user data within the type of time period to obtain the set of probabilities of the corresponding age groups of the type of time period; If there are type-two time periods with a data volume equal to 1 among the various time periods, then determine the age group probability set of the user data within the type-two time periods as the age group probability set of the type-two time periods; If there are type-three time periods with a data volume less than 1 among the various time periods, then determine the age group probability set of the type-three time periods according to the age group probability set of the previous adjacent time period of the type-three time periods.

5. The method according to claim 1, wherein Calculate the relevant parameter set of each time period according to the age group probability set of each time period, including: Calculate the score of each time period according to the age group probability set of each time period, and determine the ranking of each time period according to the score of each time period, so as to obtain the relevant parameter set of each time period.

6. The method according to claim 5, wherein Determine the ranking of each time period according to the score of each time period, including: Determine the scores of other users in each time period; For the same time period, determine the ranking of the target user among the other users according to the scores of the other users and the score of the target user, and thus obtain the ranking of the target user in each time period.

7. The method according to claim 5, wherein Calculate the score of each time period according to the age group probability set of each time period, including: Perform probability fusion within the set of the age group probability sets of each time period to obtain the reference probability corresponding to each time period; Calculate the score of each time period according to the reference probability and the age group probability set corresponding to each time period.

8. The method according to claim 7, wherein Calculate the score of each time period according to the reference probability and the age group probability set corresponding to each time period, including: Calculate the probability mean and probability standard deviation of each time period according to the reference probability of each time period; Calculate the score of each time period according to the reference probability, probability mean and probability standard deviation of each time period.

9. The method according to claim 1, wherein Generate user features according to the age group probability set of each time period and the relevant parameter set of each time period, including: Generate the time parameter corresponding to each piece of user data according to the time stamp of each piece of user data; wherein, the time parameter is used to represent whether the corresponding user data is generated on a weekday; Generate the user features according to the time parameter corresponding to each piece of user data, the age group probability set of each time period and the relevant parameter set of each time period.

10. The method according to claim 9, characterized in that, Generate the user features according to the time parameter corresponding to each piece of user data, the age group probability set of each time period and the relevant parameter set of each time period, including: Determine the time parameter of each time period according to the time parameter corresponding to each piece of user data; Generate the user features according to the time parameter of each time period, the age group probability set of each time period and the relevant parameter set of each time period.

11. The method according to claim 10, wherein Generate the user features according to the time parameter of each time period, the age group probability set of each time period and the relevant parameter set of each time period, including: Concatenate the time parameters within each of the time periods, the probability sets of the corresponding age groups for each of the time periods, and the rankings in the set of relevant parameters for each of the time periods to obtain the sub-features corresponding to each of the time periods; Concatenate the sub-features corresponding to each of the time periods in time series to obtain the user features.

12. The method according to claim 1, wherein Based on the second prediction model, determine the determination basis corresponding to the user features, and perform specific user identification on the target user according to the determination basis, including: Input the user features into the second prediction model to calculate the probability of the specific user corresponding to the user features through the second prediction model as the determination basis; If the probability of the specific user is within the preset threshold range, determine that the target user is a specific user.

13. An artificial intelligence-based data processing device, characterized in that, Including: A parameter prediction unit for obtaining each piece of user data of the target user and predicting the probability sets of the corresponding age groups for each piece of user data based on the first prediction model; A parameter aggregation unit for reading the timestamps in each piece of user data and aggregating each piece of user data into each time period of the time series according to the timestamps of each piece of user data to obtain an aggregation result; A parameter determination unit for determining the probability sets of the corresponding age groups for each time period according to the aggregation result and the probability sets of the corresponding age groups for each piece of user data; A parameter calculation unit for calculating the set of relevant parameters for each time period according to the probability sets of the corresponding age groups for each time period; wherein, any set of relevant parameters in the set of relevant parameters for each time period includes: a score and a ranking, the score is used as the personalized representation of each time period, and the ranking is used to represent the ranking order of the score of the target user in the time period compared with the scores of other users in the time period; A feature generation unit for generating user features according to the probability sets of the corresponding age groups for each time period and the set of relevant parameters for each time period; A specific user determination unit for determining the determination basis corresponding to the user features based on the second prediction model, and performing specific user identification on the target user according to the determination basis, where the first prediction model and the second prediction model correspond to different model parameters.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1-12.

15. An electronic device, characterized in that, Including: A processor; And A memory for storing the executable instructions of the processor; Wherein, the processor is configured to execute the method according to any one of claims 1-12 by executing the executable instructions.

16. A computer program product, characterized in that, The computer program product includes computer instructions, the computer instructions are stored in a computer-readable storage medium, the processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method according to any one of claims 1-12 above.

Images