Package recommendation method, control device, and storage medium

Abstract

The application discloses a package recommendation method, a control device and a storage medium. The method comprises the following steps: constructing a first user vector based on a first consumption amount of each product service of a target user; determining a second consumption amount of each product service of a neighbor user based on second historical consumption information, second historical package prices and all product services of the neighbor user, and constructing a second user vector; updating the first user vector based on the second user vector and the number of communications to obtain a target user vector; and recommending a target package to the target user based on a first similarity between the target user vector and each reference vector in a reference package set. After obtaining the first user vector based on the quota of each product service of the target user, the first user vector is updated based on the second user vector, the similarity with each reference package is calculated, and the target package is recommended. The package consumption and the social relationship are comprehensively considered, and the accuracy of the recommended result is improved.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of data processing technology, and in particular to a package recommendation method, control device, and storage medium. Background Technology

[0002] Telecommunications operators regularly release new telecommunications packages to meet diverse user needs. When a user wants to change their current package, the operator prioritizes recommending new packages. Current recommendation methods only consider the user's data usage and historical package specifications. They recommend packages based on similarity between the candidate packages in the new package set and the user's data usage and historical package specifications. However, user package selection is often influenced by social circles (such as recommendations from family, friends, and colleagues). Recommendations based solely on similarity to historical package usage or specifications fail to capture these socially driven selection patterns, leading to package recommendations that deviate from actual needs and resulting in a poor user experience. Summary of the Invention

[0003] This application provides a package recommendation method, control device, and storage medium, which enables the package recommendation results to better meet the user's actual needs and improve the user experience.

[0004] In a first aspect, embodiments of this application provide a method for recommending service packages, applied to a telecommunications user management system, wherein the telecommunications user management system has multiple pre-set product services, and the method includes:

[0005] When a target user's request to change their service plan is received, a set of reference plans is obtained. The set of reference plans includes multiple different reference plans, each of which corresponds to a reference vector. Each reference vector consists of multiple credit limits, the number of which is the same as the number of product services. Each credit limit is the amount spent on the corresponding product service for the corresponding reference plan's service plan specifications.

[0006] Based on the target user's first historical usage information, first historical package price, and all the product services, the target user's first consumption amount for each product service is determined, and a first user vector of the target user is constructed based on each first consumption amount, wherein the higher the value, the higher the target user's preference value for the product service corresponding to the first consumption amount;

[0007] Based on the second historical usage information, second historical package price, and all the product services of each neighboring user who has a communication relationship with the target user, the second consumption amount of the neighboring user for each of the product services is determined, and a second user vector of the neighboring user is constructed based on each of the second consumption amounts;

[0008] The first user vector is iteratively updated based on the second user vector and the corresponding number of communications to obtain the target user vector.

[0009] Calculate a first similarity between the target user vector and each of the reference vectors, select a target package from the set of reference packages based on the first similarity of each reference package, and recommend the target package to the target user.

[0010] In some embodiments, the first user vector is iteratively updated based on the second user vector and the corresponding number of communications to obtain the target user vector, including:

[0011] The weight of each neighboring user is calculated based on the number of neighboring users and the number of communications between each neighboring user and the target user.

[0012] Based on the first user vector, all the second user vectors, and their corresponding weights, the first user vector is iteratively updated until the absolute value of the difference between the user vector obtained in the previous iteration and the user vector corresponding to the current iteration is less than a preset value. The user vector corresponding to the current iteration is then determined as the target user vector.

[0013] In some embodiments, the weight of each neighboring user is calculated based on the number of neighboring users and the number of communications between each neighboring user and the target user, according to the following formula:

[0014]

[0015] Among them, w k The weight is k, where k is the sequence number of the neighboring user, and Count is the number of the neighboring user. k Let k be the number of communications between the neighboring user corresponding to the serial number k and the target user, and M be the number of the neighboring users corresponding to the target user.

[0016] In some embodiments, the first user vector is iteratively updated based on the first user vector, all of the second user vectors, and the corresponding weights, according to the following formula:

[0017]

[0018] Among them, Vu n+1 Vu is the updated first user vector corresponding to the target user. n The first user vector before the update, where n=1, Vu1 is the first user vector, k is the index of the neighboring user, and Count is the number of the neighboring user. kLet k be the number of communications between the neighboring user corresponding to the sequence number k and the target user, M be the number of the neighboring users corresponding to the target user, and λ be the learning rate hyperparameter.

[0019] In some embodiments, obtaining the reference package set includes:

[0020] Obtain an initial set of packages, wherein each candidate package in the initial set of packages has a corresponding package type tag;

[0021] Based on the reference type tag of the target user's current package and the historical package attribute information, candidate tags are selected from all the package type tags, wherein the candidate tags are tags that match the keywords associated with the reference type tag of the target user's current package and the historical package attribute information;

[0022] The reference package set is formed based on the candidate packages corresponding to the candidate tags.

[0023] In some embodiments, a first similarity between the target user vector and each of the reference vectors is calculated according to the following formula:

[0024] Simi(user,plan) = μ*Simi(Vu last V Plan )+(1-μ)*Simi(Spec user ,Spec plan );

[0025] Where Simi(user,plan) is the first similarity, μ is a preset hyperparameter, and Vu last Let V be the target user vector. Plan For any of the aforementioned reference vectors, Spec user The Spec represents the current plan specifications for the target user. plan For any reference package's package specifications, Simi(x, y) is used to calculate the similarity between x and y.

[0026] In some embodiments, the number of target packages is multiple. Selecting a target package from the set of reference packages based on the first similarity of each reference package, and recommending the target package to the target user, includes:

[0027] Based on the first similarity corresponding to each of the reference packages, the reference packages are reordered in descending order of their numerical values;

[0028] A predetermined number of packages ranked at the top of the sorted reference packages are selected as target packages, and a recommendation list is formed based on all the target packages.

[0029] The recommended list is sent to the target user, or the order of the target packages in the recommended list is adjusted based on preset popular package elements, and the adjusted recommended list is sent to the target user.

[0030] In some embodiments, adjusting the order of the target packages in the recommendation list based on preset popular package elements, and sending the adjusted recommendation list to the target user, includes:

[0031] Determine the attribute content of each product in each of the target packages, and calculate the second similarity between the popular package elements and the attribute content;

[0032] The target packages in the recommendation list are sorted in descending order of the second similarity, and the sorted recommendation list is sent to the target user.

[0033] Secondly, embodiments of this application provide a control device, including at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor, which, when executed by the at least one control processor, enable the at least one control processor to perform the package recommendation method as described in the first aspect.

[0034] Thirdly, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for performing the package recommendation method as described in the first aspect.

[0035] This application provides a package recommendation method, control device, and storage medium. The method includes: upon receiving a package change request from a target user, obtaining a set of reference packages, wherein the set of reference packages includes multiple different reference packages, each reference package corresponds to a reference vector, and the reference vector consists of multiple credit limits, the number of credit limits being the same as the number of product services, and each credit limit being the amount spent on the corresponding product service for the package specifications of the corresponding reference package; determining the target user's first consumption amount for each of the product services based on the target user's first historical usage information, first historical package price, and all the product services; and constructing a first user vector for the target user based on each of the first consumption amounts, wherein the numerical values... The higher the first consumption amount, the higher the preference value of the target user for the corresponding product service. Based on the second historical usage information, second historical package price, and all the product services of each neighboring user with whom the target user has a communication relationship, the second consumption amount of each neighboring user for each product service is determined. A second user vector of each neighboring user is constructed based on each second consumption amount. The first user vector is iteratively updated based on the second user vector and the corresponding number of communications to obtain the target user vector. A first similarity between the target user vector and each reference vector is calculated. Based on the first similarity of each reference package, a target package is selected from the reference package set and recommended to the target user. According to the scheme provided in this application embodiment, the amount for each product service is determined based on the target user's historical package information to form a first user vector. After updating the first user vector based on the second user vectors of neighboring users with whom the target user has a communication relationship, a target package is recommended to the target user based on the similarity between the updated first user vector and each reference package in the reference package set. In this way, the influence of social relationships on the user's package selection can be considered comprehensively while also taking into account historical package usage, making the package recommendation results more in line with the user's actual needs and improving the user experience. Attached Figure Description

[0036] Figure 1 This is a flowchart of the steps of a package recommendation method provided in one embodiment of this application;

[0037] Figure 2 This is a structural diagram of a control device provided in another embodiment of this application. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0039] It is understandable that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, or the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0040] Telecommunications operators regularly release new telecommunications packages to meet diverse user needs. When a user wants to change their current package, the operator prioritizes recommending new packages. Current recommendation methods only consider the user's data usage and historical package specifications. They recommend packages based on similarity between the candidate packages in the new package set and the user's data usage and historical package specifications. However, user package selection is often influenced by social circles (such as recommendations from family, friends, and colleagues). Recommendations based solely on similarity to historical package usage or specifications fail to capture these socially driven selection patterns, leading to package recommendations that deviate from actual needs and resulting in a poor user experience.

[0041] To address the aforementioned problems, this application provides a package recommendation method, control device, and storage medium. The method includes: upon receiving a target user's package change request, obtaining a set of reference packages, wherein the set of reference packages includes multiple different reference packages, each reference package corresponds to a reference vector, and the reference vector consists of multiple credit limits, the number of credit limits being the same as the number of product services, and each credit limit being the amount spent on the corresponding product service for the corresponding package specification of the reference package; determining the target user's first consumption amount for each of the product services based on the target user's first historical usage information, first historical package price, and all the product services; and constructing a first user vector for the target user based on each of the first consumption amounts. The higher the numerical value, the higher the preference value of the target user for the product service corresponding to the first consumption amount. Based on the second historical usage information, second historical package price, and all the product services of each neighboring user with communication relationships with the target user, the second consumption amount of each neighboring user for each product service is determined. A second user vector of the neighboring user is constructed based on each second consumption amount. The first user vector is iteratively updated based on the second user vector and the corresponding number of communications to obtain the target user vector. A first similarity between the target user vector and each reference vector is calculated. Based on the first similarity of each reference package, a target package is selected from the reference package set and recommended to the target user. According to the scheme provided in this application embodiment, the amount for each product service is determined based on the target user's historical package information to form a first user vector. After updating the first user vector based on the second user vectors of neighboring users with communication relationships with the target user, a target package is recommended to the target user based on the similarity between the updated first user vector and each reference package in the reference package set. In this way, the influence of social relationships on the user's package selection can be considered comprehensively while also considering historical package usage, making the package recommendation result more in line with the user's real needs and improving the user experience.

[0042] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0043] refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of a package recommendation method provided in one embodiment of this application. This embodiment of the application provides a package recommendation method applied to a telecommunications user management system. The telecommunications user management system has multiple pre-set product services, and the method includes, but is not limited to, the following steps:

[0044] Step S10: When a target user's package change request is received, a set of reference packages is obtained. The set of reference packages includes multiple different reference packages. Each reference package has a corresponding reference vector. Each reference vector consists of multiple credit limits. The number of credit limits is the same as the number of product services. Each credit limit is the amount spent on the corresponding product service for the package specifications of the corresponding reference package.

[0045] Specifically, in this embodiment, the reference package set is a set of reference packages selected by the telecommunications user management system based on the user attributes of the target user.

[0046] Specifically, in some embodiments, Figure 1 Step S10, obtaining the reference package set, includes, but is not limited to, the following steps:

[0047] Step S11: Obtain the initial package set. Each candidate package in the initial package set has a corresponding package type label.

[0048] Step S12: Based on the reference type tag of the target user's current package and the historical package attribute information, select candidate tags from all package type tags, wherein the candidate tags are tags that match the keywords associated with the reference type tag of the target user's current package and the historical package attribute information.

[0049] Step S13: Form a reference package set based on the candidate packages corresponding to the candidate tags.

[0050] It is understandable that since the telecommunications user management system has hundreds or thousands of pre-set telecommunications packages, iterating through each pre-set package based on the characteristics of each target user who needs to change packages would consume a lot of system resources. The filtering method lacks specificity, affecting the efficiency of package recommendations and consequently impacting user experience. Therefore, this embodiment uses the reference type tags of the target user's current package and historical package attribute information. Upon receiving a target user's package change request, this embodiment first determines the target user's user profile in the system based on the reference type tags of the target user's current package and historical package attribute information. For example, it categorizes users by group (individual or family) and consumption behavior (low-demand or medium-to-high data usage). Since each candidate package in the system has a package type tag containing keywords related to the corresponding candidate package content, keyword matching is performed between the package type tags and the target user's current package reference type tags and historical package attribute information. Successfully matched tags are identified as candidate tags, and a reference package set is formed based on the candidate packages corresponding to these tags. This achieves rapid initial filtering and provides effective support for improving the efficiency of subsequent package recommendations.

[0051] Step S20: Based on the target user's first historical usage information, first historical package price, and all product services, determine the target user's first consumption amount for each product service. Construct the target user's first user vector based on each first consumption amount, where the higher the value, the higher the target user's preference value for the product service corresponding to the first consumption amount.

[0052] Step S30: Based on the second historical usage information, second historical package price and all product services of each neighboring user who has a communication relationship with the target user, determine the second consumption amount of each neighboring user for each product service, and construct the second user vector of each neighboring user based on the second consumption amount.

[0053] It is understood that this embodiment constructs a first user vector for the target user and second user vectors for each neighboring user with communication relationships with the target user. The first user vector Vu = (offer1, offer2, ..., offerN) corresponds to the target user's total consumption amount across all products and services in the system, where offer1...offerN represent the target user's current package consumption amount across different products and services. The first user vector is positively correlated with the target user's usage across all products and services in the system. The second user vector Vu... k = (offer1, offer2, ..., offerN) corresponds to the consumption amount of neighboring users for all products and services in the system. The second user vector is positively correlated with the usage of neighboring users for all products and services in the system. In this way, the correlation between the target user and the usage of neighboring users for all products and services is determined.

[0054] Step S40: Iteratively update the first user vector based on the second user vector and the corresponding number of communications to obtain the target user vector.

[0055] It is understood that, referring to the description of the above embodiments, after obtaining the first user vector and the second user vector, this embodiment iteratively updates the first user vector based on the second user vector and the corresponding number of communications to obtain the target user vector. The obtained target user vector can combine the target user's usage behavior of system products and services with the target user's social relationships to comprehensively judge the user's preference for different products. That is, it can comprehensively consider the historical package usage while also considering the impact of social relationships on the user's choice of package, providing an effective data foundation for obtaining package recommendation results that meet the user's needs.

[0056] Specifically, in some embodiments, Figure 1 Step S40 includes, but is not limited to, the following steps:

[0057] Step S41: Calculate the weight of each neighboring user based on the number of neighboring users and the number of communications between each neighboring user and the target user;

[0058] Step S42: Based on the first user vector, all the second user vectors and their corresponding weights, iteratively update the first user vector until the absolute value of the difference between the user vector obtained in the previous iteration and the user vector corresponding to the current iteration is less than a preset value, and determine the user vector corresponding to the current iteration as the target user vector.

[0059] It should be noted that in this embodiment, the weight of each neighboring user is calculated based on the number of neighboring users and the number of communications between each neighboring user and the target user, according to the following formula:

[0060]

[0061] Among them, w k The weight is k, where k is the index of the neighboring user, and Count is the number of the neighboring user. k Let k be the number of communications between the neighboring user corresponding to the sequence number k and the target user, and M be the number of neighboring users corresponding to the target user.

[0062] It should be noted that in this embodiment, the first user vector is iteratively updated based on the first user vector, all the second user vectors, and their corresponding weights, as obtained by the following formula:

[0063]

[0064] Among them, Vu n+1 Vu is the first user vector corresponding to the updated target user. n Vu1 is the first user vector before the update. When n=1, Vu1 is the first user vector, k is the index of the neighboring user, and Count is the first user vector. k Let k be the number of communications between the neighboring user corresponding to index k and the target user, M be the number of neighboring users corresponding to the target user, and λ be the learning rate hyperparameter.

[0065] It should be noted that the iteration termination condition for the iterative update operation in this embodiment is shown in the following formula:

[0066] |Vu n+1 -Vu n |<ε;

[0067] Where ε is a preset value.

[0068] It is understood that the preset value in this embodiment is a very small value. That is, during the iterative update of the first user vector based on the second user vector and the corresponding number of communications, preference information can be transmitted by utilizing the communication relationship between neighboring users and the target user to obtain a target user vector that comprehensively considers historical package usage while also considering the package preference of the target user driven by social relationships. This continues until the absolute value of the difference between the first user vector before and after the update is less than the preset value (the specific size of the preset value is not limited, and can be determined by those skilled in the art according to the actual situation). This indicates that the influence of the second user vector of the current neighboring user on the first user vector of the target user has reached a balanced state. Determining the first user vector at this time as the target user vector can ensure the stability of the target user vector, thereby ensuring the accuracy of the subsequent determination of the target package.

[0069] Step S50: Calculate the first similarity between the target user vector and each reference vector, select the target package from the reference package set based on the first similarity of each reference package, and recommend the target package to the target user.

[0070] Specifically, in some embodiments, Figure 1 Step S50, which involves selecting a target package from the reference package set based on the first similarity of each reference package and recommending the target package to the target user, includes, but is not limited to, the following steps:

[0071] Step S51: Based on the first similarity corresponding to each reference package, reorder the reference packages in descending order of their values;

[0072] Step S52: Select a preset number of top-ranked packages from the sorted reference packages and determine them as target packages; then form a recommendation list based on all target packages.

[0073] Step S53: Send the recommendation list to the target user, or adjust the order of the target packages in the recommendation list based on preset popular package elements, and send the adjusted recommendation list to the target user.

[0074] It should be noted that the first similarity score in this embodiment is calculated using the following method:

[0075] Simi(user,plan) = μ*Simi(Vu last V Plan )+(1-μ)*Simi(Spec user ,Spec plan );

[0076] Where Simi(user,plan) is the first similarity, μ is a preset hyperparameter, and Vu lastV is the target user vector. Plan For any reference vector, Spec user For the target user's current plan specifications (such as data allowance, call duration, and other basic attributes), Spec plan For any reference package, Simi(x, y) is used to calculate the similarity between x and y. It should be noted that the embodiments of this application do not limit the specific method of calculating the similarity Simi(x, y), which can be represented by calculating the cosine distance or Euclidean distance between the reference vector and the target user vector.

[0077] Among them, Spec user This is obtained by normalizing the target user vector and multiplying the normalized result by the target user's latest average monthly spending. This indicates the allocation of the target user's current plan's total price across different product services. Similarly, Spec... plan This includes the credit limit for all products and services under any reference package (used to indicate the allocation of credit limit across different products and services for any reference package's total price), i.e., Simi (Spec). user ,Spec plan μ represents the similarity of package preferences, Simi(Vu). last V Plan Similarity to package specifications (Simi) user ,Spec plan The relative importance of these factors. That is to say, the first similarity in this embodiment takes into account the similarity of package preferences and the similarity of package specifications. The similarity result is reliable and effectively ensures the reliability of the subsequent selection of the target package.

[0078] Understandably, in this embodiment, based on the first similarity score corresponding to each reference package, the reference packages are reordered in descending order of value. A preset number of packages at the top of the sorted list are selected as target packages. Thus, the determined target packages are closer to the preferences and package specifications of the target user vector. A recommendation list is formed based on all target packages and sent to the target user. Alternatively, the order of the target packages in the recommendation list can be adjusted based on preset popular package elements, and the adjusted recommendation list is sent to the target user. In this way, the recommendation list sent to the target user can better reflect the user's actual needs.

[0079] Specifically, in some embodiments, step S53, which involves adjusting the order of target packages in the recommendation list based on preset popular package elements and sending the adjusted recommendation list to the target user, includes, but is not limited to, the following steps:

[0080] Step S531: Determine the attribute content of each product in each target package, and calculate the second similarity between popular package elements and attribute content;

[0081] Step S532: Adjust the order of the target packages in the recommendation list according to the second similarity from high to low, and send the adjusted recommendation list to the target user.

[0082] Specifically, in this embodiment, the popular package elements are keywords of products and services that are currently highly accepted by users.

[0083] It is understandable that individual user behavior data may lag behind market changes, while popular package elements implicitly contain group preferences and operator strategies. The unadjusted recommendation list already contains recommendations that match users' actual needs. Based on this, the attribute content of each product in each target package in the recommendation list is determined, the second similarity between popular package elements and attribute content is calculated, and the ranking of target packages in the recommendation list is adjusted according to the second similarity from high to low. The adjusted recommendation list is then sent to the target users. In other words, the recommendation list adjusted based on the current popular package elements can dynamically adapt to market trends and user group preferences, thus better optimizing the user experience.

[0084] like Figure 2 As shown, Figure 2 This is a structural diagram of a control device provided in one embodiment of this application. The present invention also provides a control device 200, comprising:

[0085] The processor 210 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0086] The memory 220 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 220 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 220 and is called and executed by the processor 210 using the package recommendation method of the embodiments of this application.

[0087] Input / output interface 230 is used to implement information input and output;

[0088] The communication interface 240 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0089] Bus 250 transmits information between various components of the device (e.g., processor 210, memory 220, input / output interface 230, and communication interface 240);

[0090] The processor 210, memory 220, input / output interface 230 and communication interface 240 are connected to each other within the device via bus 250.

[0091] In addition, this application embodiment also provides a storage medium, which is a computer-readable storage medium, storing a computer program that, when executed by a processor, implements the above-described package recommendation method.

[0092] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof. The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separate, and may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0093] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0094] The above provides a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A method for recommending meal packages, characterized in that, Applied to a telecommunications user management system, which has multiple pre-set product services, the method includes: When a target user's request to change their service plan is received, a set of reference plans is obtained. The set of reference plans includes multiple different reference plans, each of which corresponds to a reference vector. Each reference vector consists of multiple credit limits, the number of which is the same as the number of product services. Each credit limit is the amount spent on the corresponding product service for the corresponding reference plan's service plan specifications. Based on the target user's first historical usage information, first historical package price, and all the product services, the target user's first consumption amount for each product service is determined, and a first user vector of the target user is constructed based on each first consumption amount, wherein the higher the value, the higher the target user's preference value for the product service corresponding to the first consumption amount; Based on the second historical usage information, second historical package price, and all the product services of each neighboring user who has a communication relationship with the target user, the second consumption amount of the neighboring user for each of the product services is determined, and a second user vector of the neighboring user is constructed based on each of the second consumption amounts; The weight of each neighboring user is calculated based on the number of neighboring users and the number of communications between each neighboring user and the target user. The first user vector is iteratively updated based on the first user vector, all the second user vectors, and their corresponding weights until the absolute value of the difference between the user vector obtained in the previous iteration and the user vector corresponding to the current iteration is less than a preset value. The user vector corresponding to the current iteration is then determined as the target user vector. The first similarity between the target user vector and each of the reference vectors is calculated. Based on the first similarity of each of the reference packages, a target package is selected from the set of reference packages, and the target package is recommended to the target user. The weight of each neighbor user is calculated based on the number of neighbor users and the number of communications between each neighbor user and the target user, according to the following formula: ； in, For the weight, k The serial number of the neighboring user. For serial number k The corresponding number of communications between the neighboring user and the target user. The number of neighboring users corresponding to the target user; The first user vector is iteratively updated based on the first user vector, all the second user vectors, and their corresponding weights, according to the following formula: ； in, The updated first user vector corresponding to the target user. The first user vector before the update, when n=1, , k The serial number of the neighboring user. For serial number k The corresponding number of communications between the neighboring user and the target user. The number of neighboring users corresponding to the target user. The learning rate is a hyperparameter; The first similarity between the target user vector and each of the reference vectors is calculated according to the following formula: in, For the first similarity, For preset hyperparameters, The target user vector, Let be any of the aforementioned reference vectors. The current plan specifications for the target user. For any reference package, the package specifications are as follows. Used to calculate the similarity between x and y; The collection of reference packages includes: Obtain an initial set of packages, wherein each candidate package in the initial set of packages has a corresponding package type tag; Based on the reference type tag of the target user's current package and the historical package attribute information, candidate tags are selected from all the package type tags, wherein the candidate tags are tags that match the keywords associated with the reference type tag of the target user's current package and the historical package attribute information; The reference package set is formed based on the candidate packages corresponding to the candidate tags.

2. The package recommendation method according to claim 1, characterized in that, Based on the first similarity of each of the reference packages, a target package is selected from the set of reference packages, and the target package is recommended to the target user, including: Based on the first similarity corresponding to each of the reference packages, the reference packages are reordered in descending order of their numerical values; Select a preset number of the top-ranked packages from the sorted reference packages to determine the target packages, and form a recommendation list based on all the target packages; The recommended list is sent to the target user, or the order of the target packages in the recommended list is adjusted based on preset popular package elements, and the adjusted recommended list is sent to the target user.

3. The package recommendation method according to claim 2, characterized in that, Adjusting the order of target packages in the recommended list based on preset popular package elements, and sending the adjusted recommended list to the target user, includes: Determine the attribute content of each product in each of the target packages, and calculate the second similarity between the popular package elements and the attribute content; The target packages in the recommendation list are sorted in descending order of the second similarity, and the sorted recommendation list is sent to the target user.

4. A control device, characterized in that, It includes at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor to enable the at least one control processor to perform the package recommendation method as described in any one of claims 1 to 3.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the package recommendation method as described in any one of claims 1 to 3.

Images