Data recommendation method and device, electronic equipment, and computer readable storage medium

Abstract

The application provides a data recommendation method and device, electronic equipment, computer program product and computer readable storage medium; the method comprises the following steps: obtaining the retention rate data of different types of objects using a target object; obtaining the update time of the target object characteristic update, and obtaining a plurality of time intervals with each update time as a node; for each type of object, the following processing is performed: curve fitting processing is performed on the retention rate data of the current type of object to obtain the retention rate curve of each time interval; based on each retention rate curve of the current type of object, the predicted life cycle index of the current type of object is determined; the predicted life cycle index corresponding to each type is sorted in descending order, and the terminal equipment of the type of object located at the head of the descending order sorting list is sent the recommendation information associated with the target object. Through the application, the required computing resources of the recommendation information can be saved, and the accuracy of information recommendation can be improved.

Description

Technical Field

[0001] This application relates to computer technology, and more particularly to a data recommendation method, apparatus, electronic device, and computer-readable storage medium. Background Technology

[0002] Most methods for determining the lifecycle of objects in related technologies rely on summarizing object behavior. When product information is diverse and the types and number of objects are extensive, determining the corresponding recommendation information for each object wastes a lot of computing resources.

[0003] Currently, there is no good way to save computational resources for recommendation information and improve its accuracy. Summary of the Invention

[0004] This application provides a data recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product, which can save the computing resources required for recommendation information and improve the accuracy of information recommendation.

[0005] The technical solution of this application embodiment is implemented as follows: This application provides a data recommendation method, the method comprising: Acquire retention rate data for different types of objects using the target item, wherein the retention rate data includes the retention rate of different types of objects at different times; Obtain the update time of the target item's characteristics, and use each update time as a node to obtain multiple time intervals; Perform the following processing for each type of object: Using each time interval as a unit, curve fitting is performed on the retention rate data of the current type of object to obtain the retention rate curve corresponding to each time interval, wherein the current type is the type currently being processed; Based on each retention rate curve corresponding to the current type of object, determine the predicted lifecycle index of the current type of object; The predicted lifecycle indicators corresponding to each type are sorted in descending order, and the recommendation information associated with the target item is sent to the terminal device of the type located at the head of the descending sort list.

[0006] This application provides a data recommendation device, including: The data acquisition module is configured to acquire retention rate data of different types of objects using the target item, wherein the retention rate data includes the retention rate of different types of objects at different times; The data acquisition module is configured to acquire the update time of the target item's characteristic update, and to acquire multiple time intervals using each update time as a node; The data processing module is configured to perform the following processing for each type of object: perform curve fitting processing on the retention rate data of the current type of object in each time interval to obtain the retention rate curve corresponding to each time interval, wherein the current type is the type currently being processed; and determine the predicted lifecycle index of the current type of object based on each retention rate curve corresponding to the current type of object. The data recommendation module is configured to sort the predicted lifecycle indicators corresponding to each type in descending order, and send the recommendation information associated with the target item to the terminal device of the type located at the head of the descending sort list.

[0007] This application provides an electronic device, the electronic device comprising: Memory is used to store executable instructions for a computer; The processor, when executing computer-executable instructions stored in the memory, implements the data recommendation method provided in the embodiments of this application.

[0008] This application provides a computer-readable storage medium storing computer-executable instructions for implementing the data recommendation method provided in this application when executed by a processor.

[0009] This application provides a computer program product, including a computer program or computer executable instructions, which, when executed by a processor, implement the data recommendation method provided in this application.

[0010] The embodiments of this application have the following beneficial effects: By characterizing the features of different objects using lifecycle metrics, information recommendation can be performed in batches for different types of objects, improving the efficiency of information recommendation and saving computing resources for recommending information to a large number of terminal devices. Based on the lifecycle metrics, the information is sorted in descending order, and recommendation information is sent to each terminal device of the type of object at the head of the descending sort list, improving the accuracy of batch information recommendation and saving computing resources of the recommendation server. Attached Figure Description

[0011] Figure 1 This is a schematic diagram illustrating the application mode of the data recommendation method provided in the embodiments of this application; Figure 2 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application; Figures 3A to 3GThis is a flowchart illustrating the data recommendation method provided in an embodiment of this application; Figures 4A to 4B This is a flowchart illustrating the data recommendation method provided in an embodiment of this application; Figure 5A This is a schematic diagram of the object retention rate function fitting; Figure 5B This is a schematic diagram of the retention rate curve fitting of the data recommendation method provided in the embodiments of this application; Figure 6 This is an optional flowchart illustrating the data recommendation method provided in the embodiments of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0014] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0015] It should be noted that in the embodiments of this application, user information, user feedback data and other related data are involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0017] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0018] 1) Items. Items can be virtual or physical. Examples of virtual items include game items, applications, and websites. Examples of physical items include home appliances and terminal devices.

[0019] 2) Lifecycle: The length of time from the creation of an item to its demise. In this embodiment, it refers to the length of time from when an object starts using the item until the object no longer uses the item. For example, the length of time from when a user starts using application A until the user stops using application A is the user's lifecycle for application A.

[0020] 3) Retention refers to the continuation of something. In the embodiments of this application, retention refers to the behavior of an object continuously using a certain item.

[0021] 4) Retention rate, calculated on a device basis. The retention rate for a given day is the ratio of the number of active retained devices to the number of newly added retained devices. For example, assuming the item is an application, the retention rate on day N is the ratio of the number of newly added terminal devices that have installed the application to the number of terminal devices that have launched the application on day N within the statistical period (week / month). N is a positive integer.

[0022] 5) Average Revenue Per User (ARPU) refers to the average amount of money each user spends on items over a period of time, such as the average daily spending per user.

[0023] 6) Lifetime Value (LTV): The total amount of money an object spends on an item over its lifetime. It can be represented as the product of average spending per user and lifetime value.

[0024] 7) Lifecycle stages are the intervals in the process of an object using an item. According to the time sequence, lifecycle stages include introduction stage, growth stage, maturity stage, dormancy stage, and churn stage.

[0025] This application provides a data recommendation method, a data recommendation device, an electronic device, a computer-readable storage medium, and a computer program product, which can save the computing resources required for recommendation information and improve the accuracy of information recommendation.

[0026] The following describes exemplary applications of the electronic devices provided in the embodiments of this application. These electronic devices can be implemented as various types of user terminals, such as laptops, tablets, desktop computers, set-top boxes, mobile devices (e.g., mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), and in-vehicle terminals, or as servers. Exemplary applications when the device is implemented as a terminal device or a server will be described below.

[0027] refer to Figure 1 , Figure 1 This is a schematic diagram illustrating the application mode of the data recommendation method provided in the embodiments of this application; for example, Figure 1 The system involves a recommendation server 200, a network 300, terminal devices 400, and a database 500. Terminal devices 400 are connected to recommendation server 200 via network 300, which can be a wide area network (WAN), a local area network (LAN), or a combination of both.

[0028] For example, the object can be a user, and the item can be a virtual or physical item; for instance, an item could be an application. Terminal device 400 is a terminal device with the application installed, and multiple terminal devices 400 may exist with the application installed. Recommendation server 200 is the advertising platform's server used to recommend item-related information to users' terminal devices 400, such as recommendations for purchasing application membership services or application version updates. Database 500 is the cloud database corresponding to the application, used to store records of terminal device logins and application usage.

[0029] In some embodiments, the recommendation server 200 obtains application-specific retention rate data for different types of objects from the database 500, obtains predicted lifecycle metrics for different types of objects based on the retention rate data, and the predicted lifecycle metrics can characterize the total time that the object uses the application. The server then sends corresponding recommendation information to each terminal device 400 of the type of object located at the head of the descending sorted list of the predicted lifecycle metrics, thereby realizing batch push of recommendation information, improving recommendation efficiency and accuracy, and saving computational resources for recommending information to multiple terminal devices.

[0030] This application embodiment can be implemented using blockchain technology. The image processing model trained in this application embodiment can be uploaded to the blockchain for storage, and the reliability of the image processing model can be guaranteed through a consensus algorithm. Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer.

[0031] This application embodiment can be implemented using database technology. A database, simply put, can be viewed as an electronic filing cabinet storing electronic files, where users can perform operations such as adding, querying, updating, and deleting data. A "database" is a collection of data stored together in a certain way, capable of being shared by multiple users, having minimal redundancy, and being independent of application programs.

[0032] A Database Management System (DBMS) is a computer software system designed to manage databases, generally possessing basic functions such as storage, retrieval, security, and backup. DBMSs can be classified according to the database model they support, such as relational or XML (Extensible Markup Language); or according to the type of computer they support, such as server clusters or mobile devices; or according to the query language used, such as Structured Query Language (SQL) or XQuery; or according to performance priorities, such as maximum scale or maximum operating speed; or other classification methods. Regardless of the classification method used, some DBMSs can cross categories, for example, simultaneously supporting multiple query languages.

[0033] This application embodiment can also be implemented through artificial intelligence (AI). AI is the theory, methods, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess perception, reasoning, and decision-making capabilities.

[0034] This application embodiment can also be implemented using cloud technology. Cloud technology is a general term for network technology, information technology, integration technology, management platform technology, and application technology based on cloud computing business models. It can form a resource pool, available on demand, offering flexibility and convenience. Cloud computing technology will become a crucial support. Backend services of technical network systems require substantial computing and storage resources, such as video websites, image websites, and many portal websites. With the rapid development and application of the internet industry, and driven by demands for search services, social networks, mobile commerce, and open collaboration, every item may eventually possess its own hash-coded identification mark, requiring transmission to a backend system for logical processing. Data at different levels will be processed separately, and various industry data will require robust system support, which can only be achieved through cloud computing.

[0035] In some embodiments, the recommended server 200 may be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Electronic devices may be smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, etc., but are not limited to these. Terminal devices and servers can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment of the invention.

[0036] See Figure 2 , Figure 2 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. The electronic device may be a recommendation server 200. Figure 2 The recommendation server 200 shown includes at least one processor 410, memory 450, and at least one network interface 420. The various components in the recommendation server 200 are coupled together via a bus system 440. It is understood that the bus system 440 is used to implement communication between these components. In addition to a data bus, the bus system 440 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 440.

[0037] Processor 410 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0038] The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 450 may optionally include one or more storage devices physically located away from the processor 410.

[0039] The memory 450 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 450 described in this application embodiment is intended to include any suitable type of memory.

[0040] In some embodiments, memory 450 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0041] Operating system 451 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks; The network communication module 452 is used to reach other electronic devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc. In some embodiments, the apparatus provided in this application can be implemented in software. Figure 2 A data recommendation device 455 stored in memory 450 is shown. This device can be software in the form of programs and plug-ins, and includes the following software modules: a data acquisition module 4551, a data processing module 4552, and a data recommendation module 4553. These modules are logically connected and can therefore be arbitrarily combined or further divided according to their implemented functions. The functions of each module will be described below.

[0042] The data recommendation method provided in the embodiments of this application will be described below. As mentioned above, the electronic device implementing the data recommendation method of the embodiments of this application can be a terminal, a server, or a combination of both. Therefore, the executing entity of each step will not be described again below.

[0043] See Figure 3A , Figure 3A This is a flowchart illustrating the data recommendation method provided in the embodiments of this application, which will be combined with... Figure 3A The steps shown are explained.

[0044] In step 301, retention rate data of different types of objects using the target item are obtained.

[0045] For example, the target item can be a virtual item or a physical item. Virtual items could be applications, games, virtual props, etc. Physical items could be electronic devices, home appliances, etc. The type of object can be distinguished based on its permission level or the information channel through which it receives recommendation information. For example, if the item is an application, the object type can be distinguished based on the permission level of the object's account within the application. Each type of object can be considered as a group of objects, with different permission levels for different types. Another example is that the channels through which the object's terminal device receives recommendation information include instant messaging software, social media software, and SMS. The object type can be distinguished based on the information channels the object commonly uses. Each type of object can be considered as a group of objects, with different information channels used for different types of objects.

[0046] For example, retention rate data includes retention rates for different types of objects at different times. Retention refers to the behavior of users continuing to use a particular item, and the retention rate for a given day is the ratio of the number of active retained devices to the number of newly retained devices.

[0047] In some embodiments, reference Figure 3B , Figure 3B This is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 301 can be achieved through the following steps 3011 to 3013, which will be described in detail below.

[0048] Perform steps 3011 through 3013 for different types of objects.

[0049] In step 3011, the first number of new objects for the target item is obtained.

[0050] For example, the new object is the object that starts using the target item; the object can be a user, and assuming the item is an application, get the first number A of new users A for the application on date X.

[0051] In step 3012, a second number of retained objects using the target item within each preset period is obtained.

[0052] For example, the retained objects are those who continue to use the target item; continuing from the example above, the second number Bi of retained users who still use the application each day after the statistical date X, and the retention rate on day n is the ratio between the number of retained users Bn on day n and the first number A.

[0053] In step 3013, a third ratio between each second quantity and the first quantity is obtained, and multiple third ratios are used as retention rate data.

[0054] For example, this retrieves the retention rate data for a specific object type. Retention rate data is data specific to a group of objects.

[0055] Continue to refer to Figure 3A In step 302, the update time of the target item's characteristics is obtained, and multiple time intervals are obtained with each update time as a node.

[0056] For example, if the target item is a virtual item, such as an application, updates to its features could include adding new functions, redesigning the interface, or introducing new promotional activities. If the target item is a physical item, such as a mobile phone, updates to its features could include system updates.

[0057] For example, let's take an application as an example. Suppose the application went online at time T1, the current time is T4, and the application's update times include T2 and T3. Then, each update time is treated as a node, dividing the time into time intervals [T1,T2], (T2,T3], and (T3,T4).

[0058] As explained above, there are multiple different types of objects that use the target item. Steps 303 and 304 are performed for each type of object.

[0059] In step 303, curve fitting is performed on the retention rate data of the current type of object for each time interval to obtain the retention rate curve corresponding to each time interval.

[0060] For example, the current type refers to the type currently being processed. Curve fitting of retention rate data for different types of objects can be performed synchronously or sequentially. Since object retention rate data is acquired periodically, it is discrete data. Curve fitting is the process of connecting each data point of this discrete data to form a line segment and fitting it to a curve. The discrete retention rate data for each time interval is fitted separately to obtain the retention rate curve for each time interval. In the coordinate system of the retention rate curve, the horizontal axis corresponds to time, and the vertical axis corresponds to retention rate.

[0061] In step 304, based on each retention rate curve corresponding to the current type of object, the predicted lifecycle index of the current type of object is determined.

[0062] For example, the predicted lifecycle metric can be predicted lifecycle (LT) or predicted lifecycle value (LTV). Predicted lifecycle characterizes the length of time an object may use a target item. Predicted lifecycle value characterizes the total amount of money an object may spend on a target item.

[0063] In some embodiments, reference Figure 3E , Figure 3E This is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 304 can be implemented through the following steps 3041E to 3042E, which are described in detail below.

[0064] In step 3041E, the following processing is performed for each retention rate curve corresponding to the current type of object: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the sub-predicted life cycle index corresponding to the time interval.

[0065] For example, using the predicted lifespan as the predicted lifespan metric, definite integral processing is used to obtain the area between each retention rate curve and the horizontal axis and the edge of the time interval. This area is also the sub-predicted lifespan metric, and the unit of this area is time, such as days.

[0066] In step 3042E, each sub-predicted lifecycle index is summed to obtain the predicted lifecycle index for the current type of object.

[0067] For example, the sub-predictive lifecycle metrics corresponding to each time interval are summed, and the sum is used as the predictive lifecycle metric.

[0068] In this embodiment, by obtaining the predicted lifecycle and making information recommendations based on the predicted lifecycle, the computing resources of the recommendation server are saved and the efficiency of sending recommendation information to terminal devices of multiple objects in batches is improved.

[0069] In some embodiments, the predicted lifetime value is used as the predicted lifetime metric, and the retention rate curve includes: a first retention rate curve before the update time and a second retention rate curve after the update time; (Refer to...) Figure 3F , Figure 3F This is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 304 can be implemented through steps 3041F to 3043F, which are described in detail below.

[0070] In step 3041F, the time interval corresponding to the first retention rate curve is set to the entire life cycle to obtain the third retention rate curve.

[0071] refer to Figure 5B , Figure 5B This is a schematic diagram of the retention rate curve fitting of the data recommendation method provided in this application embodiment; wherein, the date corresponding to the dashed line T represents the date of the update of the target item's characteristics. The second retention rate curve is the dashed curve 502B. Assuming that the first retention rate curve is curve 501B, if the time interval corresponding to the first retention rate curve is set to the entire life cycle, then the third retention rate curve (a combination of curves 501B and 503B) can be obtained.

[0072] In step 3042F, the first ratio between the following retention rates is obtained: the retention rate corresponding to the second retention rate curve and the retention rate corresponding to the third retention rate curve at each time point.

[0073] For example, before the update time, 1 is used as the first ratio. After the update time, the retention rate corresponding to the second retention rate curve and the retention rate corresponding to the third retention rate curve are obtained for each time step.

[0074] In step 3043F, a first sequence is formed based on each retention rate corresponding to the third retention rate curve.

[0075] For example, suppose the retention rates corresponding to the third retention rate curve are respectively to They can form the first sequence. .

[0076] In step 3044F, a second sequence is formed based on the first ratio, and the first sequence is multiplied by the second sequence to obtain the retention rate matrix.

[0077] For example, suppose day t is the update time. The first ratio before day t is 1, and the first ratio on day t is... The second sequence can be obtained. .

[0078] Example, retention rate matrix This can be expressed as the following formula (3):

[0079] in, It is a function of the lifetime retention rate on day t. It is a power function representing the retention rate of objects, where a and b are constants. It is the ratio between the retention rate after the new feature is applied on day t and the retention rate that is not affected.

[0080] Retention rate matrix Expressed in vector form as the following formula (4):

[0081] In step 3045F, the predicted average spending amount for the target item for the current type of object is obtained.

[0082] For example, the predicted average spending is the average amount each object spends on the target item.

[0083] In some embodiments, step 3045F can be implemented as follows: The ratio of the following parameters is used as the current average spending amount for the current type of object: the total spending amount for the target item by the current type of object, the number of objects of the current type, and the duration of use of the target item by the current type of object; the difference between the ratio of the following parameters and 1 is used as a first weight value: the second spending amount corresponding to the target item after its characteristic update, and the first spending amount corresponding to the target item before its characteristic update. The product of the sum of 1 and the first weight value and the current average spending amount is used as the predicted average spending amount.

[0084] For example, the second spending limit is reduced compared to the first spending limit; the second spending limit can be the price corresponding to the discount on the target item, and the first spending limit is the price before the discount.

[0085] Example, current average spending amount =Total spending amount / (Number of objects of the current type) (Duration of time the target item is used by objects of the current type); First weight value = (Second spending amount / First spending amount) - 1; Predicted average spending amount = Current average spending amount (First weight value +1).

[0086] In some embodiments, the type of an object corresponds to the information channel used by the object; step 3045F can be implemented in the following way: the ratio of the following parameters is used as the current average consumption amount of the current type of object: the total consumption amount of the current type of object for the target item, the number of the current type of object, and the duration of the current type of object using the target item; the first proportion of objects with different permission levels in the information channel is obtained; for each permission level of object, the second ratio between the following parameters is obtained: the second consumption amount corresponding to the target item after the feature update, and the first consumption amount corresponding to the target item before the feature update; the first proportion of the object for each permission level is multiplied by the second ratio, and the sum of the results of each multiplication is used as the second weight value; the product of the second weight value and the current average consumption amount is used as the predicted average consumption amount.

[0087] For example, the second spending limit is reduced compared to the first spending limit; for instance, an update to the target item's features is a discount for application membership privileges. Then the second ratio for VIP members... = (Discounted VIP membership amount / Original VIP membership amount). This calculates the percentage of each membership level, such as the percentage of regular members, VIPs, and SVIPs. , and Then the second weight value of the target item. .

[0088] If the weight is 1 before the feature update, then the second weight value W is expressed as the following formula (5):

[0089] In step 3046F, the predicted average spending amount is multiplied by the retention rate matrix to obtain the predicted lifecycle index.

[0090] For example, by combining the above-mentioned spending weight vector with the current average spending per object, the updated average spending per object over the object's lifecycle can be obtained. It can be expressed as the following formula (6):

[0091] In this embodiment, the lifetime value is obtained, and information recommendation is performed based on the predicted lifetime value, which saves the computing resources of the recommendation server and improves the efficiency of sending recommendation information to terminal devices of multiple objects in batches.

[0092] In some embodiments, reference Figure 3G , Figure 3GThis is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 304 can be implemented through the following steps 3041G to 3043G, which are described in detail below.

[0093] In step 3041G, the following processing is performed for each retention rate curve corresponding to the current type of object: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the definite integral processing result of the time interval.

[0094] For example, the principle of step 3041G is the same as that of step 3041E, and will not be repeated here.

[0095] In step 3042G, the results of each definite integral are summed to obtain the total definite integral result.

[0096] For example, the principle of step 3042G is the same as that of step 3042E, and will not be repeated here.

[0097] In step 3043G, the predicted average spending amount for the target item for the current type of object is obtained.

[0098] For example, the principle of step 3043G is the same as that of step 3045F above, and will not be repeated here.

[0099] In step 3044G, the predicted average consumption amount is multiplied by the total fixed integral processing result to obtain the predicted life cycle index.

[0100] For example, lifecycle value can be represented as the product of predicted lifecycle and predicted average consumption amount, and the result of the total definite integral processing above is also the predicted lifecycle.

[0101] In this embodiment, the lifetime value is obtained, and information recommendation is performed based on the predicted lifetime value, which saves the computing resources of the recommendation server and improves the efficiency of sending recommendation information to terminal devices of multiple objects in batches.

[0102] Continue to refer to Figure 3A In step 305, the predicted lifecycle indicators corresponding to each type are sorted in descending order, and the recommended information associated with the target item is sent to the terminal device of the type at the head of the descending sort list.

[0103] For example, the recommended information associated with the target item could be advertisements for the target item, advertisements for alternative products of the target item, or update information about the target item.

[0104] For example, the type at the head of the descending sorted list, that is, the type at the top of the descending sorted list with a preset proportion or a preset quantity. For instance, if there are 10 different types in the descending sorted list and the preset proportion is 10%, then the first type at the top of the descending sorted list is selected as the target type. Another example: if the preset quantity is 2, then the first and second types at the top of the descending sorted list are selected as the target types.

[0105] In some embodiments, the type of an object corresponds to the information channel used by the object; see reference. Figure 3C , Figure 3C This is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 305 can be implemented through steps 3051C to 3053C, as detailed below.

[0106] In step 3051C, the predicted lifecycle indicators corresponding to each information channel are sorted in descending order to obtain the first descending sorted list.

[0107] For example, an information channel is the source of information. The source of recommended information could be a short video platform, SMS notifications, or the official WeChat account of an instant messaging software. The information channel to which an object belongs can be distinguished by the account used to log in to the application. For example, an object can log in to the application using a social media account, an instant messaging software account, or a mobile phone number (corresponding to SMS). The predicted lifecycle indicators for each information channel are sorted in descending order to obtain a first descending sorted list, for example: Predicted lifecycle indicator 1 - Short video platform, Predicted lifecycle indicator 2 - Social media platform, Predicted lifecycle indicator 3 - Official WeChat account of instant messaging software, Predicted lifecycle indicator 4 - SMS. Among these, predicted lifecycle indicators 1 through 4 decrease sequentially.

[0108] In step 3052C, at least one information channel located at the head of the first descending sorted list is selected as the target information channel.

[0109] For example, continuing from the example above, if we select the two information channels at the head of the first descending sorted list as the target information channels, then the short video platform and the WeChat official account of the instant messaging software will be selected as the target information channels.

[0110] In step 3053C, each terminal device of the object corresponding to the target information channel is obtained, and recommendation information associated with the target item is sent to each terminal device based on the target information channel.

[0111] For example, the system obtains the terminal devices of objects using short video platforms and instant messaging software, and sends recommendation information related to the target items to these terminal devices based on the official accounts of the short video platforms and instant messaging software.

[0112] This application improves the efficiency of information recommendation by recommending information based on object type according to lifecycle indicators, thus avoiding the computational resource consumption caused by generating and recommending information for each object separately.

[0113] In some embodiments, the type of an object corresponds to the object's permission level; see reference. Figure 3D , Figure 3D This is a flowchart illustrating the data recommendation method provided in this application embodiment. Step 305 can be implemented through steps 3051D to 3053D, as detailed below.

[0114] In step 3051D, the predicted lifecycle indicators corresponding to each permission level are sorted in descending order to obtain a second descending sorted list.

[0115] For example, the permission level of an object can be a membership level in an application, such as: Regular Member, VIP Member, and Super VIP Member (SVIP Member). The predicted lifecycle indicators for each membership level are sorted in descending order to obtain a second descending list, for example: Predicted Lifecycle Indicator A - Regular Member, Predicted Lifecycle Indicator B - Super VIP Member, Predicted Lifecycle Indicator C - VIP Member. The predicted lifecycle indicators A through C decrease sequentially.

[0116] In step 3052D, at least one permission level located at the head of the second descending sorted list is taken as the target permission level.

[0117] For example, continuing from the example above, if we select the permission level at the head of the second descending sorted list as the target permission level, then the ordinary member is the target permission level.

[0118] In step 3053D, each terminal device corresponding to the target permission level is obtained, and recommendation information associated with the target item is sent to each terminal device.

[0119] For example, the terminal devices that acquire objects of the ordinary member category are retrieved, and recommendation information associated with the target item is sent to these terminal devices.

[0120] This application improves the efficiency of information recommendation by recommending information based on object type according to lifecycle indicators, thus avoiding the computational resource consumption caused by generating and recommending information for each object separately.

[0121] In some embodiments, reference Figure 4A , Figure 4A This is a flowchart illustrating the data recommendation method provided in this application embodiment; before step 305, recommendation information can be determined through the following steps 3061A to 3063A, which are described in detail below.

[0122] In step 3061A, the object of the type located at the head of the descending sorted list is taken as the target object, and the behavior data of the target object for the target item is obtained.

[0123] For example, behavior refers to actions that an object has performed. If the target item is a virtual item, the behavior includes logging in, purchasing, and using; if the item is a physical item, the behavior includes purchasing, using, repairing, and reselling.

[0124] In step 3062A, the lifecycle stage to which the target object belongs is determined based on the behavior type of the behavior data.

[0125] For example, each lifecycle stage has a pre-defined mapping relationship with different behavior types, and the current lifecycle stage of an object can be determined based on the mapping relationship.

[0126] The lifecycle stages represent the intervals in the process of using an item. According to the time sequence, the lifecycle stages include the introduction stage (e.g., corresponding to the behavior of starting to use such as purchasing or registering), the growth stage (e.g., corresponding to the behavior of using or logging in), the maturity stage (e.g., corresponding to the behavior of consuming or recommending the target item), the dormant stage (e.g., corresponding to the behavior of repair or resale), and the churn stage (corresponding to the cessation of use).

[0127] In step 3063A, the recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device.

[0128] For example, each lifecycle stage has a pre-defined mapping relationship with different types of recommendation information. For instance: the introduction stage corresponds to user manual-type recommendations, the growth and maturity stages correspond to advertising information, the dormant stage corresponds to advertising information on promotional activities for the target item, and the churn stage corresponds to advertising information on alternative products for the target item.

[0129] This application embodiment determines the recommended terminal devices to the target through different lifecycle stages, which improves the accuracy of the recommendation. It also saves the computing resources required to obtain the recommendation information by judging the type of the recommendation information.

[0130] In some embodiments, reference Figure 4B , Figure 4BThis is a flowchart illustrating the data recommendation method provided in this application embodiment; before step 305, the recommendation information to be sent can be determined through the following steps 3061B to 3065B, which are described in detail below.

[0131] In step 3061B, the object of the type located at the head of the descending sorted list is taken as the target object.

[0132] For example, the target object is a group of objects.

[0133] In step 3062B, in response to receiving the request information sent by the terminal device of the target object, the time information carried in the request information is obtained.

[0134] For example, the content carried in the request information may be the identifier of the terminal device, the identifier of the target object (such as ID), and time information, specifically the time when the request information was sent.

[0135] In step 3063B, the usage duration of the target item by the target object is determined based on time information, and a fourth ratio between the usage duration and the predicted life cycle index is obtained.

[0136] For example, the recommendation server can retrieve information from the database to determine when a target object begins using a target item, and determine the duration of the target object's use of the target item based on the time information carried in the request. When the predicted lifecycle metric is the lifecycle, a fourth ratio between the usage duration and the predicted lifecycle metric is obtained.

[0137] In step 3064B, the lifecycle stage to which the target object belongs is determined based on the fourth ratio.

[0138] For example, each lifecycle stage is mapped to a different ratio range; the ratio range corresponding to each lifecycle stage can be determined experimentally based on the application scenario.

[0139] For example, by obtaining the ratio range to which the fourth ratio belongs and determining the lifecycle stage corresponding to the ratio range based on the mapping relationship, the lifecycle stage to which the target object belongs can be obtained.

[0140] In step 3065B, the recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device.

[0141] For example, step 3065B can be performed as step 3063A, which will not be repeated here.

[0142] This application embodiment determines the recommended terminal devices to the target through different lifecycle stages, which improves the accuracy of the recommendation. It also saves the computing resources required to obtain the recommendation information by judging the type of the recommendation information.

[0143] This application embodiment uses lifecycle indicators to characterize the features of different objects, enabling batch information recommendation for different types of objects, improving the efficiency of information recommendation, and saving computing resources for recommending information to a large number of terminal devices; based on the lifecycle indicators, the recommendation information is sent to each terminal device of the type of object at the head of the descending sort list in descending order, improving the accuracy of batch information recommendation and saving computing resources of the recommendation server.

[0144] The following will describe an exemplary application of the data recommendation method of this application in a real-world application scenario.

[0145] Time series models, used in production and scientific research, involve observing and measuring one or a set of variables x(t), arranging them in chronological order across a series of times t1, t2, …, tn (t being the independent variable), and using this data to explain the variables and their relationships. The most common applications are weather forecasting and credit limit prediction. Based on time series models, the total lifetime value of an object can be calculated, and this total lifetime value can serve as reference data for recommendation information.

[0146] Ideally, when the characteristics of security products are stable and the product's spending strategy tends towards a long cycle, the consumer's spending will be stable. In estimating the total lifetime value, this application embodiment summarizes the long-term behavior of the object from its short-term behavior. Related technologies estimate the total lifetime value of the object by inputting various short-term behavior data, obtaining the average spending per user and retention rate from the lifetime estimation model, and substituting them into the lifetime value estimation formula (1) for calculation:

[0147] LTV stands for Lifetime Value. This refers to the retention rate function, where ARPU is the average spending per user. Related technologies, when considering long-term user behavior, typically use a single function to fit historical data, lacking the ability to incorporate changes in product characteristics. For example, they don't use a power function to fit the retention rate one week after a new user to obtain the full lifecycle retention rate, making them unsuitable for scenarios where user retention changes due to product characteristics.

[0148] In related technologies, behavior is grouped, and then prediction models are built for different objects to predict total lifecycle value. However, this extensive method of using the same prediction model produces significant errors. Its drawbacks include over-reliance on historical behavioral data of the objects, failure to incorporate future influencing factors, and failure to quantify and input foreseeable product plans and behavioral changes into the model.

[0149] The data recommendation method provided in this application improves the accuracy of retention rate by fitting the retention rate based on the time interval of adding new features to items, making the prediction of the total lifetime value of objects more accurate. Based on the total lifetime value, the accuracy of information recommendation for terminal devices for different types of objects can be improved.

[0150] This application's embodiments are based on a lifetime value prediction model, integrating retention rate growth sequence and credit factor to obtain a more accurate object lifetime value prediction model. See also Figure 6 , Figure 6 This is a flowchart illustrating the data recommendation method provided in the embodiments of this application, which will be combined with... Figure 6 The steps shown are explained.

[0151] In step 601, the short-term retention rate is calculated, a long-term retention power function is fitted based on the short-term retention rate, and the object retention rate is obtained based on the long-term retention power function.

[0152] In this embodiment of the application, the product can be an application, and the object is the user. The retention rate is calculated based on the terminal device with the application installed. The following is an explanation based on the above examples.

[0153] For example, in the short-term retention rate, "short-term" can refer to one day. The short-term retention rate is calculated in the following two ways: 1. Calculated on an account basis: Retention rate = Active account retention / New account retention * 100%. 2. Calculated on a device basis: Retention rate = Active device retention / New device retention * 100%.

[0154] The long-term retention rate is obtained by fitting a power function based on the short-term retention rate over multiple days. For example, the true retention rate in the early stage of the object's lifecycle (such as the first 7 days) is used to fit the long-term retention rate. The object's retention rate follows the trend of a power function. Table (1) below shows an example of fitting the object's lifecycle retention rate: Table (1)

[0155] For example, Table (1) shows the retention rates of newly added users on different dates for the next day, day 2, ..., day 7 of their lifecycle. Table (1) is a statistical analysis of short-term retention rates. Based on the short-term retention rates, the average retention rates for different dates can be obtained, as shown in Table (2) below: Table (2)

[0156] For example, the characteristic of object retention rate changes is a decaying trend, and the decay rate gradually slows down, eventually reaching a relatively stable low level. A power function can be used to fit the object's full lifecycle retention rate. (Reference) Figure 5A 5A is a schematic diagram of the object retention rate function fitting. Figure 5A The horizontal axis represents the date, and the vertical axis represents the retention rate. Each point corresponds to a retention rate in Table (2). The dashed line is the curve fitted to the retention rate function.

[0157] In step 602, the update time of the new features of the product is obtained, and the retention rate sequence of the new features of the product is obtained based on the update time.

[0158] For example, products are constantly being updated and upgraded. For instance, if the product is an application, then the application may have the following types of new features: new functions, interface redesign, new promotional activities, etc. Before launching new features, applications generally need to conduct small-scale experiments to evaluate the differences between the current version and the experimental version in various metrics. When collecting experimental data, the impact of whether or not a new feature is launched on the retention rate can be obtained, such as: retention rate increases, retention rate decreases. The magnitude of change in retention rate is dynamic. For example, if a new feature is launched on day T of the object's lifecycle, it can be represented by the following formula (2):

[0159] Where F is the new feature retention rate sequence, and 1 represents the retention rate that was not affected before the new feature was added. Characterizes the ratio between the retention rate after the new feature is applied on day T and the retention rate that is not affected.

[0160] For example, steps 602 and 601 can be executed simultaneously. Step 605 is executed after steps 602 and 601.

[0161] In step 605, the sequence of object retention rate and product new feature retention rate is merged.

[0162] For example, after obtaining the early retention rate of an object and fitting its long-term retention rate, the retention rate sequence of new product features is incorporated into the fitted power function to obtain a new lifetime retention rate function, expressed as the following formula (3):

[0163] in, It is a function of the lifetime retention rate on day t. It is a power function representing the retention rate of objects, where a and b are constants. It is the ratio between the retention rate after the new feature is applied on day t and the retention rate that is not affected.

[0164] Step 605 can be expressed in vector form as the following formula (4):

[0165] refer to Figure 5B , Figure 5B This is a schematic diagram of the retention rate curve fitting of the data recommendation method provided in the embodiments of this application. The date corresponding to the dashed line T represents the date when the product adds new features. The two sides of the dashed line T correspond to the fitting curve 501B and the fitting curve 502B affected by the new features, respectively.

[0166] In this embodiment of the application, by adding a product new feature retention rate sequence, the accuracy of obtaining the retention rate is improved, and the accuracy of fitting the retention rate curve is improved.

[0167] Continue to refer to Figure 6 In step 603, the average spending per object is calculated.

[0168] For example, retrieve the consumption records of multiple objects for the application, such as calculating the average daily consumption per object based on the early lifecycle payment data of the objects, such as the average consumption per object for the first 7 days of the object's lifecycle.

[0169] In step 604, a quota weight vector is obtained based on the weight of different levels and the proportion of objects of different levels.

[0170] For example, an object's level refers to its permission level for the product. For instance, the level can be determined by the user's membership level registered with the application. Since the application implements different discount methods for different membership levels, and calculates the weight of different types of paid objects, when setting different levels of discounts or price increases for different membership levels, it is necessary to assign corresponding weights to different levels. Assume the application has three membership levels: Regular Member, VIP Member, and Premium VIP Member.

[0171] The credit limit weight for regular members is The weighting of VIP members' spending limits is The SVIP member's credit limit is : = (Discounted Regular Membership Quota / Original Regular Membership Quota) = (VIP membership amount after discount / original VIP membership amount) = (Discounted SVIP membership amount / Original SVIP membership amount) Calculate the percentage of each membership level during the credit limit adjustment, such as the percentage of regular members, VIPs, and SVIPs. , and Therefore, the overall product credit limit weight is... If the weight before the quota is 1, then the member quota weight vector W is expressed by the following formula (5):

[0172] For example, steps 603 and 604 can be executed simultaneously. Step 606 is executed after steps 603 and 604.

[0173] In step 606, the average spending amount per object is merged with the spending amount weight vector.

[0174] For example, by combining the above-mentioned spending weight vector with the current average spending per object, the updated average spending per object over the object's lifecycle can be obtained. It can be expressed as the following formula (6):

[0175] For example, steps 605 and 606 can be executed simultaneously. Step 607 is executed after steps 605 and 606.

[0176] In step 607, the total lifecycle value prediction model is integrated.

[0177] After quantifying the weight vector of the quota and the sequence of changes in the retention rate of new product features, each parameter is integrated into the total lifetime value prediction model to obtain formula (7):

[0178] This application's embodiments effectively integrate the impact of changes in product characteristics and spending strategies on the prediction of total lifetime value. For example, if a new feature is added to a product one month after its launch, it is expected to improve user retention, which will consequently increase the user's total lifetime value. This application's embodiments make the prediction of user lifetime value more accurate by introducing product expectation influencing factors.

[0179] In some embodiments, users can obtain product information through different information channels, such as receiving product-related advertising videos from short video platforms, receiving promotional SMS messages related to product offers, or obtaining product advertising data from the official accounts of instant messaging software. For another example, information channels can be differentiated based on the account used by the user to log in to the application; the user can log in to the application using a social media account, an instant messaging software account, or a mobile phone number.

[0180] This application embodiment can be deployed on an activity operation system or a channel operation system. The underlying system development logic calculates the total lifetime value (TPV), and the results can be used to calculate channel efficiency ranking factors, predict channel operation effects, etc. It obtains the TPV corresponding to the user group for each channel, sorts the TPV of each channel in descending order, and selects the channel with the highest TPV for advertising information delivery. This allows channel placement personnel or operations personnel to select efficient and high-quality channels when planning channel placement and operational activities, improving the efficiency of batch advertising information recommendations.

[0181] This application embodiment uses lifecycle indicators to characterize the features of different objects, enabling batch information recommendation for different types of objects, improving the efficiency of information recommendation, and saving computing resources for recommending information to a large number of terminal devices; based on the lifecycle indicators, the recommendation information is sent to each terminal device of the type of object at the head of the descending sort list in descending order, improving the accuracy of batch information recommendation and saving computing resources of the recommendation server.

[0182] The following description continues to illustrate the exemplary structure of the data recommendation device 455 provided in the embodiments of this application as a software module. In some embodiments, such as Figure 2As shown, the software modules stored in the data recommendation device 455 in the memory 440 may include: a data acquisition module 4551, configured to acquire retention rate data of different types of objects using the target item, wherein the retention rate data includes the retention rate of different types of objects at different times; the data acquisition module 4551 is configured to acquire the update time of the target item's feature update, and acquire multiple time intervals with each update time as a node; a data processing module 4552, configured to perform the following processing for each type of object: perform curve fitting processing on the retention rate data of the current type of object in each time interval to obtain the retention rate curve corresponding to each time interval, wherein the current type is the type currently being processed; determine the predicted life cycle index of the current type of object based on each retention rate curve corresponding to the current type of object; and a data recommendation module 4553, configured to sort the predicted life cycle index corresponding to each type in descending order, and send the recommendation information associated with the target item to the terminal device of the type of object located at the head of the descending sort list.

[0183] In some embodiments, the type of an object corresponds to the information channel used by the object; the data recommendation module 4553 is configured to perform descending sorting on the predicted life cycle indicators corresponding to each information channel to obtain a first descending sorted list; at least one information channel located at the head of the first descending sorted list is taken as the target information channel; each terminal device of the object corresponding to the target information channel is obtained, and recommendation information associated with the target item is sent to each terminal device based on the target information channel.

[0184] In some embodiments, the type of an object corresponds to the permission level of the object; the data recommendation module 4553 is configured to sort the predicted lifecycle indicators corresponding to each permission level in descending order to obtain a second descending sort list; take at least one permission level at the head of the second descending sort list as the target permission level; obtain each terminal device of the object corresponding to the target permission level, and send the recommendation information associated with the target item to each terminal device.

[0185] In some embodiments, the data processing module 4552 is configured to perform the following processing for each retention rate curve corresponding to the current type of object: perform definite integral calculation processing based on the time interval corresponding to the retention rate curve to obtain the sub-predicted life cycle index corresponding to the time interval; and perform summation processing on each sub-predicted life cycle index to obtain the predicted life cycle index of the current type of object.

[0186] In some embodiments, the retention rate curve includes: a first retention rate curve before the update time and a second retention rate curve after the update time; the data processing module 4552 is configured to set the time interval corresponding to the first retention rate curve to the entire lifecycle to obtain a third retention rate curve; obtain a first ratio between the following retention rates: the retention rate corresponding to the second retention rate curve at each time point and the retention rate corresponding to the third retention rate curve; form a first sequence based on each retention rate corresponding to the third retention rate curve; form a second sequence based on the first ratio, and multiply the first sequence and the second sequence to obtain a retention rate matrix; obtain the predicted average spending amount of the current type of object for the target item; multiply the predicted average spending amount by the retention rate matrix to obtain a predicted lifecycle index.

[0187] In some embodiments, the data processing module 4552 is configured to perform the following processing for each retention rate curve corresponding to the current type of object: perform definite integral calculation processing based on the time interval corresponding to the retention rate curve to obtain the definite integral processing result of the time interval; add up each definite integral processing result to obtain the total definite integral processing result; obtain the predicted average consumption amount of the current type of object for the target item; and multiply the predicted average consumption amount with the total definite integral processing result to obtain the predicted lifecycle index.

[0188] In some embodiments, the data processing module 4552 is configured to use the ratio of the following parameters as the current average consumption amount of the current type of object: the total consumption amount of the current type of object for the target item, the product of the number of current type of object and the duration of use of the target item by the current type of object; use the difference between the ratio of the following parameters and 1 as a first weight value: the second consumption amount corresponding to the target item after the feature update, and the first consumption amount corresponding to the target item before the feature update, wherein the second consumption amount is reduced compared to the first consumption amount; and use the product of the sum of 1 and the first weight value and the current average consumption amount as the predicted average consumption amount.

[0189] In some embodiments, the type of an object corresponds to the information channel used by the object; the data processing module 4552 is configured to use the ratio of the following parameters as the current average consumption amount of the current type of object: the total consumption amount of the current type of object for the target item, the product of the number of the current type of object and the duration of the current type of object using the target item; the first proportion of objects with different permission levels among the objects in the information channel; for each permission level of object, obtain a second ratio between the following parameters: the second consumption amount corresponding to the target item after the feature update, and the first consumption amount corresponding to the target item before the feature update, wherein the second consumption amount is reduced compared to the first consumption amount; multiply the first proportion of objects with the second ratio for each permission level, and sum the results of each multiplication as a second weight value; and use the product of the second weight value and the current average consumption amount as the predicted average consumption amount.

[0190] In some embodiments, the data recommendation module 4553 is configured to, before sending recommendation information associated with a target item to a terminal device containing an object of a type located at the head of a descending sorted list, take the object at the head of the descending sorted list as the target object, obtain behavioral data of the target object toward the target item; determine the lifecycle stage to which the target object belongs based on the behavioral type of the behavioral data, wherein each lifecycle stage has a mapping relationship with a different behavioral type; and use the recommendation information associated with the lifecycle stage to which the target object belongs as recommendation information to send to the terminal device, wherein each lifecycle stage has a mapping relationship with a different recommendation information.

[0191] In some embodiments, the data recommendation module 4553 is configured to, before sending recommendation information associated with a target item to a terminal device containing an object of a type located at the head of a descending sorted list, designate the object at the head of the descending sorted list as the target object; in response to a request from the terminal device receiving the target object, obtain time information carried in the request information; determine the usage duration of the target item by the target object based on the time information, and obtain a fourth ratio between the usage duration and a predicted lifecycle indicator; determine the lifecycle stage to which the target object belongs based on the fourth ratio, wherein each lifecycle stage is mapped to a different ratio range; and use the recommendation information associated with the lifecycle stage to which the target object belongs as recommendation information to send to the terminal device, wherein each lifecycle stage is mapped to a different recommendation information.

[0192] In some embodiments, the data acquisition module 4551 is configured to perform the following processing for different types of objects: acquiring a first number of newly added objects for the target item, wherein the newly added objects are objects that have started using the target item; acquiring multiple second numbers of retained objects that use the target item within each preset period, wherein the retained objects are objects that continue to use the target item; acquiring a third ratio between each second number and the first number, and using the multiple third ratios as retention rate data.

[0193] This application provides a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer-executable instructions from the computer-readable storage medium and executes the computer-executable instructions, causing the electronic device to perform the data recommendation method described in this application.

[0194] This application provides a computer-readable storage medium storing computer-executable instructions. When these computer-executable instructions are executed by a processor, they cause the processor to execute the data recommendation method provided in this application, for example... Figure 3A The data recommendation method is shown.

[0195] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0196] In some embodiments, computer-executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0197] As an example, computer-executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0198] As an example, executable instructions can be deployed to execute on a single electronic device, or on multiple electronic devices located in one location, or on multiple electronic devices distributed across multiple locations and interconnected via a communication network.

[0199] In summary, this application's embodiments, by characterizing the features of different objects using lifecycle metrics, enable batch information recommendation for different types of objects, improving the efficiency of information recommendation and saving computational resources for recommending information to a large number of terminal devices. Furthermore, by sorting the objects in descending order based on lifecycle metrics and sending recommendation information to each terminal device of the type at the head of the descending sort list, the accuracy of batch information recommendation is improved, further saving computational resources of the recommendation server.

[0200] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. A data recommendation method, characterized in that, The method includes: Acquire retention rate data for multiple types of objects using the target item, wherein the retention rate data includes the retention rate of each type of object at different times; Obtain the update time of the target item's characteristics, and use each update time as a node to obtain multiple time intervals; For each of the plurality of types, perform the following processing, where the current type is the type currently being processed among the plurality of types: Using each time interval as a unit, perform curve fitting on the retention rate data of the current type of object to obtain the retention rate curve corresponding to each time interval; Based on each retention rate curve corresponding to the current type of object, determine the predicted lifecycle index of the current type of object; The multiple types are sorted in descending order of their corresponding predicted lifecycle indicators. A descending sorted list is then obtained. Recommendation information associated with the target item is sent to the terminal device of the type at the head of the descending sorted list.

2. The method according to claim 1, characterized in that, The type of the object corresponds to the information channel used by the object; The process involves sorting the multiple types according to their respective predicted lifecycle indicators from largest to smallest, obtaining a descending sorted list, and sending the recommended information associated with the target item to the terminal device of the type at the head of the descending sorted list. This includes: According to the predicted life cycle index corresponding to each information channel in descending order, the multiple information channels are sorted to obtain a first descending sorted list as the descending sorted list; At least one of the information channels located at the head of the first descending sorted list is taken as the target information channel; Obtain each terminal device corresponding to the target information channel, and send the recommendation information associated with the target item to each terminal device based on the target information channel.

3. The method according to claim 1, characterized in that, The type of the object corresponds to the permission level of the object; The process involves sorting the multiple types according to their respective predicted lifecycle indicators from largest to smallest, obtaining a descending sorted list, and sending the recommended information associated with the target item to the terminal device of the type at the head of the descending sorted list. This includes: According to the predicted lifecycle index corresponding to each permission level, the multiple permission levels are sorted in descending order to obtain a second descending sort list as the descending sort list. Take at least one of the permission levels located at the head of the second descending sorted list as the target permission level; Obtain each terminal device corresponding to the target permission level, and send the recommendation information associated with the target item to each terminal device.

4. The method according to claim 1, characterized in that, The step of determining the predicted lifecycle metric for the current type of object based on each retention rate curve corresponding to the current type of object includes: For each retention rate curve corresponding to the current type of object, the following processing is performed: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the sub-predicted lifecycle index corresponding to the time interval; Each of the sub-predicted lifecycle metrics is summed to obtain the predicted lifecycle metric for the current type of object.

5. The method according to claim 1, characterized in that, The retention rate curve includes: a first retention rate curve located before the update time and a second retention rate curve located after the update time; The step of determining the predicted lifecycle metric for the current type of object based on each retention rate curve corresponding to the current type of object includes: The time interval corresponding to the power function used to fit the first retention rate curve is extended from before the update time to the entire life cycle to obtain the third retention rate curve. For each time point after the update time, a first ratio between the following retention rates is obtained: the retention rate corresponding to the second retention rate curve at that time and the retention rate corresponding to the third retention rate curve at that time, and the first ratio for each time point before the update time is determined to be 1; The first sequence is formed based on the retention rate corresponding to each moment of the third retention rate curve throughout the entire life cycle; A second sequence is formed based on the first ratio at each moment of the entire life cycle, and the retention rate matrix is ​​obtained by multiplying the first sequence and the second sequence. Obtain the predicted average spending amount for the target item for the current type of object; The predicted average spending amount is multiplied by the retention rate matrix to obtain the predicted lifecycle index.

6. The method according to claim 1, characterized in that, The step of determining the predicted lifecycle metric for the current type of object based on each retention rate curve corresponding to the current type of object includes: For each retention rate curve corresponding to the current type of object, the following processing is performed: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the definite integral processing result of the time interval; Add up the results of each definite integral processing to obtain the total definite integral processing result; Obtain the predicted average spending amount for the target item for the current type of object; The predicted average consumption amount is multiplied by the total fixed integral processing result to obtain the predicted life cycle index.

7. The method according to claim 5 or 6, characterized in that, The step of obtaining the predicted average spending amount for the target item for the current type of object includes: The ratio between the total spending of the current type of object on the target item and the product of the number of the current type of object and the duration of time the current type of object uses the target item is used as the current average spending of the current type of object; The first weight value is the difference between the ratio of the second consumption amount corresponding to the updated characteristics of the target item and the first consumption amount corresponding to the unupdated characteristics of the target item and 1, wherein the second consumption amount is reduced compared to the first consumption amount. The product of 1 and the first weight value and the current average consumption amount is used as the predicted average consumption amount.

8. The method according to claim 5 or 6, characterized in that, The type of the object corresponds to the information channel used by the object; The step of obtaining the predicted average spending amount for the target item for the current type of object includes: The ratio between the total spending of the current type of object on the target item and the product of the number of the current type of object and the duration of time the current type of object uses the target item is used as the current average spending of the current type of object; The first proportion of objects with different permission levels among the objects that obtain the information channels; For each of the aforementioned permission levels, the ratio between the second consumption amount corresponding to the target item after the feature update and the first consumption amount corresponding to the target item before the feature update is obtained as the second ratio, wherein the second consumption amount is reduced compared to the first consumption amount. The first ratio and the second ratio of objects for each of the aforementioned permission levels are multiplied together, and the sum of the results of each multiplication is used as the second weight value. The product of the second weight value and the current average spending amount is used as the predicted average spending amount.

9. The method according to claim 1, characterized in that, Before sending the recommendation information associated with the target item to the terminal device of the type located at the head of the descending sort list, the method further includes: Take the object of type located at the head of the descending sorted list as the target object, and obtain the behavior data of the target object for the target item; The lifecycle stage to which the target object belongs is determined based on the behavior type of the behavior data, wherein each lifecycle stage is mapped to a different behavior type; The recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device, wherein each lifecycle stage is mapped to different recommendation information.

10. The method according to claim 4, characterized in that, Before sending the recommendation information associated with the target item to the terminal device of the type located at the head of the descending sort list, the method further includes: The object of the type located at the head of the descending sorted list is taken as the target object; In response to receiving a request from the terminal device of the target object, the time information carried in the request is obtained; Based on the time information, the usage duration of the target item by the target object is determined, and a fourth ratio between the usage duration and the predicted life cycle index is obtained, wherein the predicted life cycle index is a predicted life cycle that characterizes the length of time the object uses the target item; The life cycle stage to which the target object belongs is determined based on the fourth ratio, wherein each life cycle stage is mapped to a different ratio range; The recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device, wherein each lifecycle stage is mapped to different recommendation information.

11. The method according to claim 1, characterized in that, The acquisition of retention rate data for multiple types of objects using the target item includes: Perform the following processing for different types of objects: Obtain a first number of new objects for the target item, wherein the new objects are objects that begin using the target item; Obtain a second number of retained objects that use the target item within each preset period, wherein the retained objects are objects that continuously use the target item; Obtain a third ratio between each of the second quantity and the first quantity, and use multiple third ratios as the retention rate data.

12. A data recommendation device, characterized in that, The device includes: The data acquisition module is configured to acquire retention rate data of multiple types of objects using the target item, wherein the retention rate data includes the retention rate of each type of object at different times; The data acquisition module is configured to acquire the update time of the target item's characteristic update, and to acquire multiple time intervals using each update time as a node; The data processing module is configured to perform the following processing on objects of each of the plurality of types, wherein the current type is the type currently being processed among the plurality of types: performing curve fitting processing on the retention rate data of the current type of object in each time interval to obtain the retention rate curve corresponding to each time interval; and determining the predicted lifecycle index of the current type of object based on each retention rate curve corresponding to the current type of object. The data recommendation module is configured to sort the multiple types in descending order according to the predicted lifecycle indicators corresponding to each type, obtain a descending sorted list, and send the recommendation information associated with the target item to the terminal device of the type located at the head of the descending sorted list.

13. The data recommendation device according to claim 12, characterized in that, The type of the object corresponds to the information channel used by the object; the data recommendation module is further configured to: According to the predicted life cycle index corresponding to each information channel in descending order, the multiple information channels are sorted to obtain a first descending sorted list as the descending sorted list; At least one of the information channels located at the head of the first descending sorted list is taken as the target information channel; Obtain each terminal device corresponding to the target information channel, and send the recommendation information associated with the target item to each terminal device based on the target information channel.

14. The data recommendation device according to claim 12, characterized in that, The type of the object corresponds to the permission level of the object; the data recommendation module is further configured as follows: According to the predicted lifecycle index corresponding to each permission level, the multiple permission levels are sorted in descending order to obtain a second descending sort list as the descending sort list. Take at least one of the permission levels located at the head of the second descending sorted list as the target permission level; Obtain each terminal device corresponding to the target permission level, and send the recommendation information associated with the target item to each terminal device.

15. The data recommendation device according to claim 12, characterized in that, The data processing module is further configured as follows: For each retention rate curve corresponding to the current type of object, the following processing is performed: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the sub-predicted lifecycle index corresponding to the time interval; Each of the sub-predicted lifecycle metrics is summed to obtain the predicted lifecycle metric for the current type of object.

16. The data recommendation device according to claim 12, characterized in that, The retention rate curve includes: a first retention rate curve located before the update time and a second retention rate curve located after the update time; the data processing module is further configured to: The time interval corresponding to the power function used to fit the first retention rate curve is extended from before the update time to the entire life cycle to obtain the third retention rate curve. For each time point after the update time, a first ratio between the following retention rates is obtained: the retention rate corresponding to the second retention rate curve at that time and the retention rate corresponding to the third retention rate curve at that time, and the first ratio for each time point before the update time is determined to be 1; The first sequence is formed based on the retention rate corresponding to each moment of the third retention rate curve throughout the entire life cycle; A second sequence is formed based on the first ratio at each moment of the entire life cycle, and the retention rate matrix is ​​obtained by multiplying the first sequence and the second sequence. Obtain the predicted average spending amount for the target item for the current type of object; The predicted average spending amount is multiplied by the retention rate matrix to obtain the predicted lifecycle index.

17. The data recommendation device according to claim 12, characterized in that, The data processing module is further configured as follows: For each retention rate curve corresponding to the current type of object, the following processing is performed: definite integral calculation is performed based on the time interval corresponding to the retention rate curve to obtain the definite integral processing result of the time interval; Add up the results of each definite integral processing to obtain the total definite integral processing result; Obtain the predicted average spending amount for the target item for the current type of object; The predicted average consumption amount is multiplied by the total fixed integral processing result to obtain the predicted life cycle index.

18. The data recommendation device according to claim 16 or 17, characterized in that, The data processing module is further configured as follows: The ratio between the total spending of the current type of object on the target item and the product of the number of the current type of object and the duration of time the current type of object uses the target item is used as the current average spending of the current type of object; The first weight value is the difference between the ratio of the second consumption amount corresponding to the updated characteristics of the target item and the first consumption amount corresponding to the unupdated characteristics of the target item and 1, wherein the second consumption amount is reduced compared to the first consumption amount. The product of 1 and the first weight value and the current average consumption amount is used as the predicted average consumption amount.

19. The data recommendation device according to claim 16 or 17, characterized in that, The type of the object corresponds to the information channel used by the object; the data processing module is further configured to: The ratio between the total spending of the current type of object on the target item and the product of the number of the current type of object and the duration of time the current type of object uses the target item is used as the current average spending of the current type of object; The first proportion of objects with different permission levels among the objects that obtain the information channels; For each of the aforementioned permission levels, the ratio between the second consumption amount corresponding to the target item after the feature update and the first consumption amount corresponding to the target item before the feature update is obtained as the second ratio, wherein the second consumption amount is reduced compared to the first consumption amount. The first ratio and the second ratio of objects for each of the aforementioned permission levels are multiplied together, and the sum of the results of each multiplication is used as the second weight value. The product of the second weight value and the current average spending amount is used as the predicted average spending amount.

20. The data recommendation device according to claim 15, characterized in that, The data recommendation module is further configured as follows: Take the object of type located at the head of the descending sorted list as the target object, and obtain the behavior data of the target object for the target item; The lifecycle stage to which the target object belongs is determined based on the behavior type of the behavior data, wherein each lifecycle stage is mapped to a different behavior type; The recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device, wherein each lifecycle stage is mapped to different recommendation information.

21. The data recommendation device according to claim 12, characterized in that, The data recommendation module is further configured as follows: The object of the type located at the head of the descending sorted list is taken as the target object; In response to receiving a request from the terminal device of the target object, the time information carried in the request is obtained; Based on the time information, the usage duration of the target item by the target object is determined, and a fourth ratio between the usage duration and the predicted life cycle index is obtained, wherein the predicted life cycle index is a predicted life cycle that characterizes the length of time the object uses the target item; The life cycle stage to which the target object belongs is determined based on the fourth ratio, wherein each life cycle stage is mapped to a different ratio range; The recommendation information associated with the lifecycle stage to which the target object belongs is used as the recommendation information to be sent to the terminal device, wherein each lifecycle stage is mapped to different recommendation information.

22. The data recommendation device according to claim 12, characterized in that, The data acquisition module is further configured as follows: Perform the following processing for different types of objects: Obtain a first number of new objects for the target item, wherein the new objects are objects that begin using the target item; Obtain a second number of retained objects that use the target item within each preset period, wherein the retained objects are objects that continuously use the target item; Obtain a third ratio between each of the second quantity and the first quantity, and use multiple third ratios as the retention rate data.

23. An electronic device, characterized in that, The electronic device includes: Memory is used to store executable instructions for a computer; A processor, when executing computer-executable instructions stored in the memory, implements the data recommendation method according to any one of claims 1 to 11.

24. A computer-readable storage medium storing computer-executable instructions, characterized in that, When the computer-executable instructions are executed by a processor, they implement the data recommendation method according to any one of claims 1 to 11.

25. A computer program product comprising a computer program or computer-executable instructions, characterized in that, When the computer program or computer-executable instructions are executed by a processor, they implement the data recommendation method according to any one of claims 1 to 11.

Images