Activity recommendation system, method and device

Abstract

The invention discloses an activity recommendation system, method and device, which are used for solving the problem of low efficiency and accuracy of manual planning of an activity scheme, and the method comprises the steps: receiving activity recommendation request information which comprises target user identification information, target activity theme information and activity constraint information; obtaining a cue word template corresponding to the target activity theme from a cue word template library; obtaining a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined according to a feedback index of the target user on historical activity recommendation data; generating a target cue word according to the target activity theme information, the activity constraint information, the target user preference tag and the cue word template; inputting the target prompt word into the large language model to obtain a target activity recommendation scheme corresponding to the target user; and displaying the target activity recommendation scheme. While the activity scheme planning efficiency is improved, the experience and feeling of the user on the activity are improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an activity recommendation system, method, and device. Background Art

[0002] Existing event plans are typically manually planned, encompassing diverse themes such as birthdays, travel, team-building, and premium experiences in the insurance industry. However, manual planning processes are time-consuming and inefficient. Furthermore, even for the same theme, plans created by different planners often vary widely, making it difficult to meet the needs of diverse users and reducing user experience.

[0003] Therefore, how to recommend activity plans more accurately to different users is one of the technical problems that need to be solved urgently by existing technologies. Summary of the Invention

[0004] In order to solve the problem of low efficiency and accuracy of manually planned activity plans, the embodiments of the present application provide an activity recommendation system, method and device.

[0005] In a first aspect, an embodiment of the present application provides an activity recommendation system, the system comprising a prompt word template library, an activity planning knowledge base, an activity recommendation module, and a display module; wherein:

[0006] The prompt word template library is used to store prompt word templates corresponding to preset activity themes;

[0007] The activity planning knowledge base is used to store the correspondence between user identification information and the user's historical activity recommendation data and user preference tags;

[0008] The activity recommendation module is configured to receive activity recommendation request information, the activity recommendation request information including target user identification information, target activity theme information, and activity constraint information; obtain a prompt word template corresponding to the target activity theme from a prompt word template library; obtain a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, the target user preference tag being determined based on a feedback indicator of the target user on historical activity recommendation data, the feedback indicator representing the target user's interest in the historical activity recommendation data; generate a target prompt word based on the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template; input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, the large language model being used to plan an activity recommendation plan for the target user based on the target prompt word;

[0009] The display module is used to display the target activity recommendation plan.

[0010] In one embodiment, the activity recommendation module is specifically used to determine the target user's feedback index on the historical activity recommendation data in the following manner: obtaining the historical activity recommendation plan of the target activity theme corresponding to the target user from the activity planning knowledge base; counting the number of exposures of the historical activity recommendation plan of the target activity theme corresponding to the target user and the number of adoptions of the historical activity recommendation plan of the target activity theme by the target user; obtaining the browsing time of the historical activity recommendation plan of the target activity theme by the target user; and determining the target user's feedback index on the historical activity recommendation data based on the number of exposures, the number of adoptions and the browsing time.

[0011] In one embodiment, the activity recommendation module is specifically configured to determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the target user's historical activity recommendation solutions for the target activity theme;

[0012] A feedback indicator of the target user on the historical activity recommendation data is determined according to the adoption rate of the target user's historical activity recommendation solutions for the target activity theme and the browsing time.

[0013] In one embodiment, the activity recommendation module is specifically configured to calculate the target user's feedback index on the historical activity recommendation data using the following formula:

[0014] W=l*logC

[0015] Wherein, W represents the target user's feedback indicator on the historical activity recommendation data;

[0016] l represents the adoption rate of the target user’s historical activity recommendation solutions for the target activity theme, m represents the number of times the target user adopts the historical activity recommendation scheme of the target activity theme, and n represents the number of times the target user is exposed to the historical activity recommendation scheme of the target activity theme;

[0017] C represents the browsing time of the target user for the historical activity recommendation solutions of the target activity theme.

[0018] In one embodiment, the activity recommendation module is specifically configured to determine the target user preference tag based on the target user's feedback indicator on historical activity recommendation data in the following manner:

[0019] If it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, extracting a first target activity object from the historical activity recommendation scheme adopted by the target user as a target user interest tag;

[0020] If it is determined that the target user's feedback index on the historical activity recommendation data is less than the preset threshold, a second target activity object is extracted from the historical activity recommendation schemes not adopted by the target user as a target user not interested label.

[0021] In one embodiment, the activity recommendation module is further configured to, if it is determined that the target activity theme includes multiple prompt word templates, determine the quality index of each prompt word template based on the historical activity recommendation data of the target activity theme for all users in the previous historical time period; and select the prompt word template with the highest quality index.

[0022] In one embodiment, the activity recommendation module is specifically configured to obtain, from the activity planning knowledge base, historical activity recommendation schemes for the target activity theme corresponding to all users within a previous historical time period; respectively count the cumulative number of exposures of the historical activity recommendation schemes for the target activity theme corresponding to each prompt word template for the target activity theme generated by all users within the previous historical time period, the cumulative number of adoptions of the historical activity recommendation schemes for the target activity theme by all users, the number of uses of the prompt word template, and the number of modifications of the prompt word template; for each prompt word template, determine, by the ratio of the cumulative number of adoptions to the cumulative number of exposures, an adoption rate of the historical activity recommendation schemes for the target activity theme generated by the prompt word template by all users; determine, by the ratio of the number of modifications of the prompt word template to the number of uses of the prompt word template, a modification degree of the prompt word template, the modification degree of the prompt word template representing the frequency of modifications to the prompt word template; and determine a quality indicator of the prompt word template based on the adoption rate of the historical activity recommendation schemes for the target activity theme by all users, the modification degree of the prompt word template, and their corresponding weights.

[0023] In one embodiment, the activity recommendation module is specifically configured to determine the quality index of the prompt word template using the following formula:

[0024] R(s)=ω1L*ω2f

[0025] Wherein, R(s) represents the quality index of the prompt word template;

[0026] L represents the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, and ω1 represents the weight of the adoption rate of the historical activity recommendation schemes of the target activity theme by all users;

[0027] f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0028] In a second aspect, an embodiment of the present application provides an activity recommendation method, comprising:

[0029] receiving an activity recommendation request message, wherein the activity recommendation request message includes target user identification information, target activity theme information, and activity constraint information;

[0030] Obtaining a prompt word template corresponding to the target activity theme from a prompt word template library;

[0031] Obtaining a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined based on a feedback indicator of the target user on the historical activity recommendation data, wherein the feedback indicator represents the target user's interest in the historical activity recommendation data;

[0032] generating a target prompt word according to the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template;

[0033] Inputting the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, wherein the large language model is used to plan an activity recommendation plan for the target user based on the target prompt word;

[0034] The target activity recommendation plan is sent to the display module for display.

[0035] In one embodiment, the target user's feedback indicator for the historical activity recommendation data is determined in the following manner:

[0036] Acquire historical activity recommendation plans for the target activity theme corresponding to the target user from the activity planning knowledge base;

[0037] Counting the number of exposures of historical activity recommendation schemes of the target activity theme corresponding to the target user and the number of adoptions of historical activity recommendation schemes of the target activity theme by the target user;

[0038] Obtaining the target user's browsing time for historical activity recommendations for the target activity theme;

[0039] A feedback index of the target user for the historical activity recommendation data is determined according to the number of exposures, the number of adoptions, and the browsing time.

[0040] In one embodiment, determining the target user's feedback indicator for the historical activity recommendation data based on the number of exposures, the number of adoptions, and the browsing time specifically includes:

[0041] The ratio of the number of adoptions to the number of exposures is determined as the target user's adoption rate of the historical activity recommendation schemes for the target activity theme;

[0042] A feedback indicator of the target user on the historical activity recommendation data is determined according to the adoption rate of the target user's historical activity recommendation solutions for the target activity theme and the browsing time.

[0043] In one embodiment, the target user's feedback indicator for the historical activity recommendation data is determined based on the target user's adoption rate of the historical activity recommendation solutions for the target activity theme and the browsing time, specifically including:

[0044] The target user's feedback index on the historical activity recommendation data is calculated using the following formula:

[0045] W=l*logC

[0046] Wherein, W represents the target user's feedback indicator on the historical activity recommendation data;

[0047] l represents the adoption rate of the target user’s historical activity recommendation solutions for the target activity theme, m represents the number of times the target user adopts the historical activity recommendation scheme of the target activity theme, and n represents the number of times the target user is exposed to the historical activity recommendation scheme of the target activity theme;

[0048] C represents the browsing time of the target user for the historical activity recommendation solutions of the target activity theme.

[0049] In one embodiment, the target user preference tag is determined according to the target user's feedback indicator on the historical activity recommendation data in the following manner:

[0050] If it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, extracting a first target activity object from the historical activity recommendation scheme adopted by the target user as a target user interest tag;

[0051] If it is determined that the target user's feedback index on the historical activity recommendation data is less than the preset threshold, a second target activity object is extracted from the historical activity recommendation schemes not adopted by the target user as a target user not interested label.

[0052] In one embodiment, the method further comprises:

[0053] If it is determined that the target activity theme includes multiple prompt word templates, determining the quality index of each prompt word template based on the historical activity recommendation data of the target activity theme for all users in the previous historical time period;

[0054] Select the prompt word template with the highest quality index.

[0055] In one embodiment, the quality index of each prompt word template is determined based on the historical activity recommendation data of the target activity theme for all users in the previous historical time period, specifically including:

[0056] Obtaining historical activity recommendation plans for the target activity theme corresponding to all users in the previous historical time period from the activity planning knowledge base;

[0057] Counting the cumulative number of exposures of historical activity recommendation schemes corresponding to the target activity theme for all users generated using each prompt word template corresponding to the target activity theme within the previous historical time period, the cumulative number of adoptions of historical activity recommendation schemes for the target activity theme by all users, the number of uses of the prompt word template, and the number of modifications of the prompt word template;

[0058] For each prompt word template, determining the ratio of the cumulative number of adoptions to the cumulative number of exposures as the adoption rate of the historical activity recommendation solutions of the target activity theme generated by the prompt word template by all users;

[0059] Determining the ratio of the number of times the prompt word template is modified to the number of times the prompt word template is used as the modification degree of the prompt word template, where the modification degree of the prompt word template represents the frequency of modification of the prompt word template;

[0060] The quality index of the prompt word template is determined according to the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, the modification degree of the prompt word template, and the corresponding weights.

[0061] In one embodiment, the quality index of the prompt word template is determined based on the adoption rate of the historical activity recommendation solutions of all users for the target activity theme, the modification degree of the prompt word template, and the corresponding weights thereof, specifically including:

[0062] The quality index of the prompt word template is determined by the following formula:

[0063] R(s)=ω1L*ω2f

[0064] Wherein, R(s) represents the quality index of the prompt word template;

[0065] L represents the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, and ω1 represents the weight of the adoption rate of the historical activity recommendation schemes of the target activity theme by all users;

[0066] f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0067] In a third aspect, an embodiment of the present application provides an activity recommendation device, comprising:

[0068] A receiving module, configured to receive activity recommendation request information, wherein the activity recommendation request information includes target user identification information, target activity theme information, and activity constraint information;

[0069] A first acquisition module is used to acquire a prompt word template corresponding to the target activity theme from a prompt word template library;

[0070] A second acquisition module is configured to acquire a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined based on a feedback index of the target user on the historical activity recommendation data, wherein the feedback index represents the target user's interest in the historical activity recommendation data;

[0071] A generating module, configured to generate a target prompt word according to the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template;

[0072] a recommendation module, configured to input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, wherein the large language model is used to plan an activity recommendation plan for the target user based on the target prompt word;

[0073] The sending module is used to send the target activity recommendation plan to the display module for display.

[0074] In one embodiment, the second acquisition module is specifically used to determine the target user's feedback index on the historical activity recommendation data in the following manner: obtaining the historical activity recommendation plan of the target activity theme corresponding to the target user from the activity planning knowledge base; counting the number of exposures of the historical activity recommendation plan of the target activity theme corresponding to the target user and the number of adoptions of the historical activity recommendation plan of the target activity theme by the target user; obtaining the browsing time of the historical activity recommendation plan of the target activity theme by the target user; and determining the target user's feedback index on the historical activity recommendation data based on the number of exposures, the number of adoptions and the browsing time.

[0075] In one embodiment, the second acquisition module is specifically used to determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the target user's historical activity recommendation schemes for the target activity theme; and determine the target user's feedback indicator for the historical activity recommendation data based on the target user's adoption rate of the historical activity recommendation schemes for the target activity theme and the browsing time.

[0076] In one embodiment, the second acquisition module is specifically configured to calculate the target user's feedback index for the historical activity recommendation data using the following formula:

[0077] W=l*logC

[0078] Wherein, W represents the target user's feedback indicator on the historical activity recommendation data;

[0079] l represents the adoption rate of the target user’s historical activity recommendation solutions for the target activity theme, m represents the number of times the target user adopts the historical activity recommendation scheme of the target activity theme, and n represents the number of times the target user is exposed to the historical activity recommendation scheme of the target activity theme;

[0080] C represents the browsing time of the target user for the historical activity recommendation solutions of the target activity theme.

[0081] In one embodiment, the second acquisition module is specifically used to determine the target user preference label based on the target user's feedback index on the historical activity recommendation data in the following manner: if it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, then extracting the first target activity object from the historical activity recommendation scheme adopted by the target user as the target user's interest label; if it is determined that the target user's feedback index on the historical activity recommendation data is less than the preset threshold, then extracting the second target activity object from the historical activity recommendation scheme not adopted by the target user as the target user's non-interest label.

[0082] In one embodiment, the device further comprises:

[0083] a determination module configured to, if it is determined that the target activity theme includes multiple prompt word templates, determine a quality indicator of each prompt word template based on historical activity recommendation data of the target activity theme for all users in a previous historical time period;

[0084] The selection module is used to select the prompt word template with the highest quality index.

[0085] In one embodiment, the determination module is specifically configured to obtain, from the activity planning knowledge base, historical activity recommendation schemes for the target activity theme corresponding to all users within a previous historical time period; respectively count the cumulative number of exposures of the historical activity recommendation schemes for the target activity theme corresponding to each prompt word template for the target activity theme generated by all users within the previous historical time period, the cumulative number of adoptions of the historical activity recommendation schemes for the target activity theme by all users, the number of uses of the prompt word template, and the number of modifications of the prompt word template; for each prompt word template, determine, by the ratio of the cumulative number of adoptions to the cumulative number of exposures, an adoption rate of the historical activity recommendation schemes for the target activity theme generated by the prompt word template by all users; determine, by the ratio of the number of modifications of the prompt word template to the number of uses of the prompt word template, a modification degree of the prompt word template, the modification degree of the prompt word template representing the frequency of modifications to the prompt word template; and determine a quality indicator of the prompt word template based on the adoption rate of the historical activity recommendation schemes for the target activity theme by all users, the modification degree of the prompt word template, and their corresponding weights.

[0086] In one embodiment, the determining module is specifically configured to determine the quality index of the prompt word template using the following formula:

[0087] R(s)=ω1L*ω2f

[0088] Wherein, R(s) represents the quality index of the prompt word template;

[0089] L represents the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, ω1 represents the weight of the adoption rate of the historical activity recommendation schemes of the target activity theme by all users; f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0090] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the activity recommendation method described in the present application when executing the program.

[0091] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the activity recommendation method described in the present application.

[0092] The beneficial effects of this application are as follows:

[0093] The activity recommendation system, method, and apparatus provided by the embodiments of the present application include a prompt word template library, an activity planning knowledge base, an activity recommendation module, and a display module. The prompt word template library is configured to store prompt word templates corresponding to preset activity themes; the activity planning knowledge base is configured to store the correspondence between user identification information and the user's historical activity recommendation data and user preference tags; the activity recommendation module is configured to receive activity recommendation request information, which includes target user identification information, target activity theme information, and activity constraint information; obtain a prompt word template corresponding to the target activity theme from the prompt word template library; and obtain a target user preference tag corresponding to the target user identification information from the activity planning knowledge base. The target user preference tag is determined based on a target user's feedback indicator on the historical activity recommendation data, which indicates the target user's interest in the historical activity recommendation data; generate a target prompt word based on the target activity theme information, activity constraint information, target user preference tag, and the prompt word template; input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user; and the large language model is configured to plan an activity recommendation plan for the target user based on the target prompt word; and the display module is configured to display the target activity recommendation plan. In an embodiment of the present application, corresponding prompt word templates are pre-set for different activity themes and stored in a prompt word template library, user feedback indicators for historical activity recommendation data are recorded in the activity planning knowledge base, and user preference labels are determined based on the user feedback indicators for historical activity recommendation data. When the activity recommendation module receives an activity recommendation request, the prompt word template corresponding to the target activity theme is obtained from the prompt word template library, and the target user preference label, target activity theme information, and activity constraint information obtained from the activity planning knowledge base are combined to generate a target prompt word. The target recommended activity plan corresponding to the target user is predicted based on the target prompt word and the large language model, and is displayed through the display module. Since the generated target prompt word is more in line with the target user's preference, while improving the efficiency of activity plan planning, the planned activity plan can be recommended to the target user more accurately, thereby enhancing the user's experience and feeling of the activity.

[0094] Other features and advantages of the present application will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present application. The purposes and other advantages of the present application can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0096] Figure 1 A schematic diagram of the structure of the activity recommendation system provided in an embodiment of the present application;

[0097] Figure 2 A flowchart of the activity recommendation method provided in an embodiment of the present application;

[0098] Figure 3 A schematic diagram of a process for determining a target user's feedback indicator for historical activity recommendation data provided in an embodiment of the present application;

[0099] Figure 4 A schematic diagram of a specific implementation process for determining target users' feedback indicators for historical activity recommendation data based on exposure times, adoption times, and browsing time provided in an embodiment of the present application;

[0100] Figure 5 A schematic diagram of a process for determining a quality indicator of a prompt word template provided in an embodiment of the present application;

[0101] Figure 6 This is a display interface diagram of the Super Body Intelligent Assistant application provided in an embodiment of the present application;

[0102] Figure 7 A schematic diagram of a target super-body activity recommendation scheme displayed in the display interface of the super-body intelligent assistant application provided in an embodiment of the present application;

[0103] Figure 8 A schematic diagram of a multi-round dialogue display interface of the Super Body Intelligent Assistant application provided in an embodiment of the present application;

[0104] Figure 9 A schematic diagram of the feedback interface of the Super Body Intelligent Assistant application provided in an embodiment of the present application;

[0105] Figure 10 A schematic diagram of the structure of the activity recommendation device provided in an embodiment of the present application;

[0106] Figure 11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0107] In order to solve the problem of low efficiency and accuracy of manually planned activity plans, the embodiments of the present application provide an activity recommendation system, method and device.

[0108] The preferred embodiments of the present application are described below in conjunction with the drawings in the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application and are not used to limit the present application. In addition, the embodiments and features in the embodiments of the present application can be combined with each other if there is no conflict.

[0109] In this article, it is necessary to understand that the technical terms involved in this application are:

[0110] An agent is generally considered to be an agent that can autonomously achieve its assigned goals through its actions. It can be manifested in software, hardware, or other physical entities. Its goal is to understand and simulate human intelligent behavior, such as learning, reasoning, decision-making, and execution. An agent can be considered a computational entity that can continuously and autonomously function and interact with its environment.

[0111] Large language model (LLM): A deep learning model that acquires extensive general knowledge based on massive text data training. It can deeply understand the meaning of text and can be used to handle complex language understanding tasks.

[0112] like Figure 1 As shown, it is a structural diagram of the activity recommendation system 10 provided in an embodiment of the present application, as shown Figure 1 The activity recommendation system shown may include: a prompt word template library 101, an activity planning knowledge base 102, an activity recommendation module 103, and a display module 104, wherein the activity recommendation module 103 and the display module 104 may be set in a computing device with computing capabilities, or the activity recommendation module 103 is set in the computing device and the display module 104 is set in the client, and the embodiment of the present application does not limit this. The prompt word module library 101 and the activity planning knowledge base 102 may be set in the computing device, and the prompt word template library 101 and the activity planning knowledge base 102 may also be set on a cloud server, and the embodiment of the present application does not limit this. When the prompt word template library 101 and the activity planning knowledge base 102 are set on the cloud server, the computing device and the cloud server are connected via a network, and the network may include, but is not limited to, a mobile communication network, a local area network, a wide area network, the Internet of Things, etc.

[0113] In the embodiment of the present application, the computing device can also be replaced by any other device with computing capabilities, such as a computer, a server, etc. The server can be an independent physical server or a cloud server that provides basic cloud computing services such as cloud servers, cloud databases, and cloud storage. The embodiment of the present application does not limit this.

[0114] The prompt word template library 101 is used to store prompt word templates corresponding to preset activity themes.

[0115] In the embodiment of the present application, the prompt word template library 101 stores prompt word templates pre-set for different activity themes, and each activity theme may include one or more prompt word templates.

[0116] The activity planning knowledge base 102 is used to store the correspondence between user identification information, user's historical activity recommendation data, and user preference tags.

[0117] In this embodiment of the present application, the activity planning knowledge base 102 stores preference tags for different users, which can be stored as a correspondence between user identification information and user preference tags. The activity planning knowledge base 102 also stores historical activity recommendation data for different users. User preference tags are determined by the activity recommendation module 103 based on user feedback indicators for the historical activity recommendation data and stored in the activity planning knowledge base 102. In other words, user preference tags are determined based on the user's interest in the historical activity recommendation data.

[0118] The activity recommendation module 103 is configured to receive activity recommendation request information, which includes target user identification information, target activity theme information, and activity constraint information; obtain a prompt word template corresponding to the target activity theme from a prompt word template library; obtain a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, where the target user preference tag is determined based on the target user's feedback index on historical activity recommendation data, which represents the target user's interest in the historical activity recommendation data; generate a target prompt word based on the target activity theme information, activity constraint information, target user preference tag, and prompt word template; input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, which is used to plan an activity recommendation plan for the target user based on the target prompt word.

[0119] In an embodiment of the present application, the activity recommendation module 103 can plan and recommend activity plans for different users on different topics. The activity recommendation module 103 includes a pre-trained large language model. An intelligent agent can also be set on the activity recommendation module 103 to run the large language model through the intelligent agent.

[0120] The display module 104 is used to display the target activity recommendation plan.

[0121] In one embodiment, the activity recommendation module is specifically used to determine the target user's feedback indicators for historical activity recommendation data in the following manner: obtaining the historical activity recommendation plans of the target activity theme corresponding to the target user from the activity planning knowledge base; counting the number of exposures of the historical activity recommendation plans of the target activity theme corresponding to the target user and the number of adoptions of the historical activity recommendation plans of the target activity theme by the target user; obtaining the target user's browsing time of the historical activity recommendation plans of the target activity theme; and determining the target user's feedback indicators for the historical activity recommendation data based on the number of exposures, the number of adoptions, and the browsing time.

[0122] In one embodiment, the activity recommendation module is specifically configured to determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the target user's historical activity recommendation solutions for the target activity theme;

[0123] Based on the target users' adoption rate and browsing time of historical activity recommendation plans for the target activity theme, determine the target users' feedback indicators for historical activity recommendation data.

[0124] In one embodiment, the activity recommendation module is specifically configured to calculate the target user's feedback index on historical activity recommendation data using the following formula:

[0125] W=l*logC

[0126] Among them, W represents the target user’s feedback indicator on historical activity recommendation data;

[0127] l represents the adoption rate of historical activity recommendations for the target activity theme by the target user, m represents the number of times the target user has adopted the historical activity recommendation schemes for the target activity theme, and n represents the number of times the target user has been exposed to the historical activity recommendation schemes for the target activity theme.

[0128] C represents the target user's browsing time for historical activity recommendations on the target activity theme.

[0129] In one embodiment, the activity recommendation module is specifically configured to determine a target user preference tag based on the target user's feedback indicators on historical activity recommendation data in the following manner:

[0130] If it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, extracting a first target activity object from the historical activity recommendation scheme adopted by the target user as a target user interest tag;

[0131] If it is determined that the target user's feedback index on the historical activity recommendation data is less than a preset threshold, a second target activity object is extracted from the historical activity recommendation schemes not adopted by the target user as a target user not interested label.

[0132] In one embodiment, the activity recommendation module is further configured to, if it is determined that there are multiple prompt word templates corresponding to the target activity theme, determine the quality index of each prompt word template based on the historical activity recommendation data of the target activity theme for all users in the previous historical time period; and select the prompt word template with the highest quality index.

[0133] In one embodiment, the activity recommendation module is specifically used to obtain historical activity recommendation plans for the target activity theme corresponding to all users in the previous historical time period from the activity planning knowledge base; respectively count the cumulative exposure times of the historical activity recommendation plans for the target activity theme corresponding to all users generated by each prompt word template corresponding to the target activity theme in the previous historical time period, the cumulative adoption times of the historical activity recommendation plans for the target activity theme by all users, the number of times the prompt word template is used, and the number of times the prompt word template is modified; for each prompt word template, the ratio of the cumulative adoption times to the cumulative exposure times is determined as the adoption rate of the historical activity recommendation plans for the target activity theme generated by the prompt word template by all users; the ratio of the number of modifications of the prompt word template to the number of times the prompt word template is used is determined as the modification degree of the prompt word template, and the modification degree of the prompt word template represents the frequency of modification of the prompt word template; and the quality index of the prompt word template is determined according to the adoption rate of the historical activity recommendation plans for the target activity theme by all users, the modification degree of the prompt word template, and the corresponding weights of each.

[0134] In one embodiment, the activity recommendation module is specifically configured to determine the quality index of the prompt word template using the following formula:

[0135] R(s)=ω1L*ω2f

[0136] Among them, R(s) represents the quality index of the prompt word template;

[0137] L represents the adoption rate of all users’ historical activity recommendation solutions for the target activity theme, and ω1 represents the weight of the adoption rate of all users’ historical activity recommendation solutions for the target activity theme;

[0138] f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0139] In an embodiment of the present application, corresponding prompt word templates are pre-set for different activity themes and stored in a prompt word template library, user feedback indicators for historical activity recommendation data are recorded in the activity planning knowledge base, and user preference labels are determined based on the user feedback indicators for historical activity recommendation data. When the activity recommendation module receives an activity recommendation request, the prompt word template corresponding to the target activity theme is obtained from the prompt word template library, and the target user preference label, target activity theme information, and activity constraint information obtained from the activity planning knowledge base are combined to generate a target prompt word. The target recommended activity plan corresponding to the target user is predicted based on the target prompt word and the large language model, and is displayed through the display module. Since the generated target prompt word is more in line with the target user's preference, while improving the efficiency of activity plan planning, the planned activity plan can be recommended to the target user more accurately, thereby enhancing the user's experience and feeling of the activity.

[0140] like Figure 2 As shown in FIG, it is a schematic diagram of the implementation flow of the activity recommendation method provided in an embodiment of the present application. The activity recommendation method can be applied to the activity recommendation module of the above-mentioned activity recommendation system, and specifically may include the following steps:

[0141] S21. Receive activity recommendation request information, where the activity recommendation request information includes target user identification information, target activity theme information, and activity constraint information.

[0142] During specific implementation, the activity recommendation module in the computing device receives the activity recommendation request information input by the user through the client or the display module in the computing device. The activity recommendation request information includes at least the target user identification information, the target activity theme information and the activity constraint information, wherein the activity constraint information is also the activity demand information, and at least includes a combination of one or more of the following information: activity name, activity type, activity budget, departure place and destination information. The target activity theme can be any theme such as birthday activity plan, travel activity plan, team building activity plan, super experience activity plan in the insurance industry, friend gathering activity plan, classmate gathering activity plan or annual meeting activity plan, etc., and the embodiment of the present application does not limit this. For example, assuming that the target activity theme is: birthday activity plan, the activity name can be: birthday party, the activity type is: immersive experience (or team collaboration), the activity budget is: within xxx yuan, the departure place is: place A, and the destination is: place B. During implementation, the activity constraint information can be flexibly set, such as other constraints, such as required elements are: scientific experiments, etc., and the embodiment of the present application does not limit this.

[0143] S22. Obtain a prompt word template corresponding to the target activity theme from a prompt word template library.

[0144] In an embodiment of the present application, the prompt word template library stores different types of activity themes and their corresponding prompt word templates. The activity themes include: birthday activity plans, travel activity plans, team building activity plans, super experience activity plans in the insurance industry, friends' gathering activity plans, classmates' gathering activity plans, annual meeting activity plans, etc. Each activity theme corresponds to at least one prompt word template, and may also include multiple prompt word templates. This embodiment of the present application does not limit this.

[0145] In specific implementation, the prompt word template corresponding to the target activity theme can be obtained through the following methods:

[0146] The target activity theme is matched with each activity theme stored in the prompt word template library. If it is determined that the target activity theme is matched, the prompt word template corresponding to the target activity theme is extracted.

[0147] The prompt word template library also includes the feature vectors of each activity theme. If it is determined that the target activity theme is not matched, the feature vector of the target activity theme is obtained according to the text embedding model, and the similarity between the feature vector of the target activity theme and the feature vector of each activity theme is calculated respectively. The prompt word template corresponding to the activity theme with the highest similarity is determined as the prompt word template corresponding to the target activity theme.

[0148] Specifically, the activity recommendation module pre-inputs each activity theme into the text embedding model to embed each activity theme, obtains the feature vector of each activity theme, and stores the correspondence between the activity theme and the feature vector of the activity theme in the prompt word template library. The text embedding model is used to generate a vector representation of the text. The text embedding model can be, but is not limited to, the following models: OpenAI's text-embedding-large-3 model, BAAI (Beijing Academy of Artificial Intelligence)'s Embedding model (such as bge-large-en-v1.5, bge-large-zh-v1.5), etc., and any other model that can generate text vectors can also be used, and the embodiments of the present application are not limited to this.

[0149] If the activity recommendation module does not match the target activity theme in the prompt word template library, it may input the target activity theme into the text embedding model to obtain a feature vector for the target activity theme. The module then calculates the similarity between the feature vector of the target activity theme and the feature vector of each activity theme stored in the prompt word template library. The activity theme corresponding to the feature vector of the activity theme with the highest similarity to the feature vector of the target activity theme is obtained, and the prompt word template corresponding to the activity theme is used as the prompt word template corresponding to the target activity theme. When calculating the similarity between the feature vector of the target activity theme and the feature vector of any activity theme in the prompt word template library, algorithms such as cosine similarity or Euclidean distance may be used, but are not limited to these algorithms. This is not a limitation of the present embodiment.

[0150] S23. Obtain the target user preference tag corresponding to the target user identification information from the activity planning knowledge base.

[0151] The activity planning knowledge base stores preference tags of different users, which can be saved as the correspondence between user identification information and user preference tags. The activity planning knowledge base also stores historical activity recommendation data of different users. User preference tags are determined based on the user's feedback index on the historical activity recommendation data. The feedback index represents the target user's interest in the historical activity recommendation data and is used to measure the target user's interest in the historical activity recommendation data.

[0152] When implementing, you can follow the following Figure 3 The process shown in FIG. 1 determines the target user's feedback indicator for the historical activity recommendation data (i.e., the target user's interest in the historical activity recommendation data), including the following steps:

[0153] S31. Obtain historical activity recommendation plans for the target activity theme corresponding to the target user from the activity planning knowledge base.

[0154] During specific implementation, the activity recommendation module obtains historical activity recommendation plans of the target activity theme corresponding to the target user from the activity planning knowledge base according to the target user identification information.

[0155] S32. Count the number of exposures of historical activity recommendation plans of the target activity theme corresponding to the target user and the number of adoptions of historical activity recommendation plans of the target activity theme by the target user.

[0156] During specific implementation, the activity recommendation module counts the number of exposures of historical activity recommendation plans of the target activity theme corresponding to the target user and the number of adoptions of historical activity recommendation plans of the target activity theme by the target user. Among them, the number of exposures of historical activity recommendation plans of the target activity theme corresponding to the target user is the number of historical activity recommendation plans of the target activity theme recommended by the activity recommendation module to the target user, and the number of adoptions of historical activity recommendation plans of the target activity theme by the target user is the number of adoptions of historical activity recommendation plans of the target activity theme recommended by the activity recommendation module by the target user.

[0157] S33: Obtain the target user's browsing time for historical activity recommendations for the target activity theme.

[0158] During specific implementation, the activity recommendation module obtains the target user's browsing time for each historical activity recommendation plan for the target activity theme recorded in the activity planning knowledge base, and calculates the average browsing time of the target user for all historical activity recommendation plans for the target activity theme as the target user's browsing time for the historical activity recommendation plans for the target activity theme.

[0159] S34. Determine the target user's feedback indicator for the historical activity recommendation data based on the number of exposures of the historical activity recommendation plans of the target activity theme corresponding to the target user, the number of adoptions of the historical activity recommendation plans of the target activity theme by the target user, and the target user's browsing time of the historical activity recommendation plans of the target activity theme.

[0160] When implementing it specifically, you can follow the following steps: Figure 4 The process shown in the figure determines the target user's feedback indicators for historical activity recommendation data based on the number of exposures, number of adoptions, and browsing time, including the following steps:

[0161] S41. Determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the historical activity recommendation plans of the target user for the target activity theme.

[0162] During specific implementation, the activity recommendation module calculates the ratio of the number of times m that the target user adopts historical activity recommendation plans for the target activity theme to the number of exposures n of historical activity recommendation plans for the target activity theme corresponding to the target user, and obtains the adoption rate l of the historical activity recommendation plans for the target activity theme by the target user.

[0163] S42. Determine the target user's feedback indicator for the historical activity recommendation data based on the target user's adoption rate and browsing time of the historical activity recommendation plan for the target activity theme.

[0164] In specific implementation, the activity recommendation module can calculate the target user's feedback index on historical activity recommendation data through the following formula:

[0165] W=l*logC

[0166] Among them, W represents the target user’s feedback indicator on historical activity recommendation data;

[0167] l represents the adoption rate of historical activity recommendations for the target activity theme by the target user, m represents the number of times the target user has adopted the historical activity recommendation schemes for the target activity theme, and n represents the number of times the target user has been exposed to the historical activity recommendation schemes for the target activity theme.

[0168] C represents the target user's browsing time for historical activity recommendations on the target activity theme.

[0169] When calculating the target user's interest in historical activity recommendation data, the original browsing time may show a power-law distribution (for example, a few historical activity recommendation solutions are viewed for a long time). Taking the logarithm of the browsing time makes the values ​​more concentrated.

[0170] Furthermore, the target user preference label can be determined based on the target user's feedback indicators on historical activity recommendation data in the following way:

[0171] If it is determined that the target user's feedback index for the historical activity recommendation data is greater than or equal to a preset threshold, the first target activity object is extracted from the historical activity recommendation scheme adopted by the target user as a target user's interest label; if it is determined that the target user's feedback index for the historical activity recommendation data is less than the preset threshold, the second target activity object is extracted from the historical activity recommendation scheme not adopted by the target user as a target user's disinterest label.

[0172] Among them, the preset threshold value can be set according to actual needs, such as being set to 0.5, which is not limited in the embodiment of the present application. When the target user's feedback index (i.e., interest level) for historical activity recommendation data is greater than or equal to the preset threshold value, it indicates that the target user is more interested in the historical activity recommendation scheme recommended by the activity recommendation module, and at least one target activity object (which can be recorded as the first target activity object) can be extracted from the historical activity recommendation scheme adopted by the target user as a target user interest tag, wherein the target activity object is an activity item in the historical activity recommendation scheme, for example, the target activity theme is a birthday activity, and an activity recommendation scheme recommended by the activity recommendation module includes the following activity items: dinner, singing birthday songs, cutting cakes, making handicrafts, visiting science theme parks, magic shows, etc., and the above at least one activity item can be extracted as a target user interest tag. When the target user's feedback index for historical activity recommendation data is less than the preset threshold value, it indicates that the target user is not very interested in the historical activity recommendation scheme recommended by the activity recommendation module, and at least one target activity object (which can be recorded as the second target activity object) can be extracted from the historical activity recommendation scheme not adopted by the target user as a target user disinterest tag.

[0173] In one embodiment, one or more activity items with the highest number in historical activity recommendation schemes adopted by the target user may be used as the first target activity object, and one or more activity items with the highest number in historical activity recommendation schemes not adopted by the target user may be used as the second target activity object.

[0174] S24: Generate target prompt words according to the target activity theme information, activity constraint information, target user preference tags and prompt word templates.

[0175] During specific implementation, the activity recommendation module adds the target activity theme information, activity constraint information, and target user preference tags to the prompt word template corresponding to the target activity theme, and generates target prompt words to assist the large language model in generating target activity recommendation plans corresponding to the target user.

[0176] If this activity recommendation request for the target activity theme is the target user's first request for the target activity theme, the activity recommendation module can return the preset preference label corresponding to the target activity theme to the display module in the client or computing device for the user to select through the display interface corresponding to the display module in the client or computing device, and then the display module in the client or computing device returns the user-selected preference label to the activity recommendation module, which fills in the prompt word template. The user can also manually fill in the preference label through the display interface corresponding to the display module in the client or computing device, and the display module sends it to the activity recommendation module, which fills in the prompt word template. The embodiment of the present application is not limited to this.

[0177] Here is a prompt word template for a birthday event theme. The template content is as follows:

[0178] “Your task is to generate an activity recommendation plan based on the provided activity requirements.

[0179] Activity requirements:

[0180] Theme positioning: [Activity theme]

[0181] Departure Place: [Departure Place]

[0182] Destination: [destination]

[0183] Activity duration: XXX hours

[0184] Number of participants: XXX

[0185] Budget range: XXX yuan (suggested allocation ratio including materials, catering, venue, etc.)

[0186] Required elements: [such as scientific experiments, role-playing, handcraft workshops, etc.]

[0187] Tags of interest: [can include multiple]

[0188] Not interested in tags: [can contain multiple]

[0189] Safety requirements: [such as age-appropriate facilities, food allergy warnings, etc.]

[0190] The activity recommendation module automatically fills the target activity theme information, activity constraint information and the interested and uninterested tags in the target user preference tags into the corresponding positions of the prompt word template to obtain the target prompt word.

[0191] If it is determined that there are multiple prompt word templates corresponding to the target activity theme, the quality index of each prompt word template can be determined based on the historical activity recommendation data of the target activity theme of all users in the previous historical time period, and the prompt word template with the highest quality index can be selected. The quality index of the prompt word template is used to measure the quality of the prompt word template. The higher the quality index value of the prompt word template, the higher the quality of the prompt word template, the more accurate the activity recommendation plan generated based on the prompt word template, and the higher the probability of being adopted by the user.

[0192] Here is a more detailed birthday event theme prompt template, the template content is as follows:

[0193] “Your task is to generate an activity recommendation plan based on the provided activity requirements.

[0194] Activity requirements:

[0195] Theme positioning: [Activity theme]

[0196] Departure Place: [Departure Place]

[0197] Destination: [destination]

[0198] Activity duration: XXX hours

[0199] Time accuracy: 15 minutes per unit

[0200] Number of participants: XXX

[0201] Personnel division of labor: [positions of host, security officer, photographer]

[0202] Budget range: XXX yuan (suggested allocation ratio including materials, catering, venue, etc.)

[0203] Required elements: [such as scientific experiments, role-playing, handcraft workshops, etc.]

[0204] Tags of interest: [can include multiple]

[0205] Not interested in tags: [can contain multiple]

[0206] Safety requirements: [such as age-appropriate facilities, food allergy warnings, etc.]

[0207] Emergency Plan: List three common ways to deal with accidents (such as conflicts between children, damage to props, etc.).

[0208] Compared with the previous prompt word template, this prompt word template further refines the time accuracy of the activity, personnel division of labor and emergency plan.

[0209] Here is another example of a birthday event theme prompt template, the template content is as follows:

[0210] “Your task is to generate an activity recommendation plan based on the provided activity requirements.

[0211] Activity requirements:

[0212] Theme positioning: [Activity theme]

[0213] Departure Place: [Departure Place]

[0214] Destination: [destination]

[0215] Activity duration: XXX hours

[0216] Number of participants: XXX

[0217] Budget range: XXX yuan (total budget controlled within XXX yuan

[0218] 70% of the materials are modified from existing household items

[0219] Contains 2 zero-cost game ideas

[0220] The dining plan focuses on homemade snacks

[0221] Provide 3 kinds of photo background DIY methods)

[0222] Required elements: [such as scientific experiments, role-playing, handcraft workshops, etc.]

[0223] Tags of interest: [can include multiple]

[0224] Not interested in tags: [can contain multiple]

[0225] Safety requirements: [such as age-appropriate facilities, food allergy warnings, etc.]

[0226] In the example of the prompt word template for this birthday event theme, the budget content is further refined compared to the first prompt word template.

[0227] When implementing it specifically, you can follow the following steps: Figure 5 The process shown determines the quality index of each prompt word template, including the following steps:

[0228] S51. Obtain historical activity recommendation plans for target activity themes corresponding to all users in the previous historical time period from an activity planning knowledge base.

[0229] During implementation, the activity recommendation module obtains historical activity recommendation plans for the target activity theme corresponding to all users within the previous historical time period from the activity planning knowledge base. The time period can be set as needed, such as to one week or half a month, or to any other duration, and this embodiment of the application does not limit this. When the activity recommendation module plans historical activity recommendation plans for the target activity theme for all users within the previous historical time period, it can use different prompt word templates corresponding to the target activity theme.

[0230] S52. Count the cumulative number of exposures of historical activity recommendation plans corresponding to the target activity theme for all users generated by using each prompt word template corresponding to the target activity theme in the previous historical time period, the cumulative number of adoptions of historical activity recommendation plans for the target activity theme by all users, the number of times the prompt word template was used, and the number of times the prompt word template was modified.

[0231] During specific implementation, the activity recommendation module counts, for each prompt word template, the cumulative number of exposures of historical activity recommendation plans corresponding to the target activity theme for all users generated by the prompt word template corresponding to the target activity theme in the previous historical time period, the cumulative number of adoptions of historical activity recommendation plans for the target activity theme by all users, the number of times the prompt word template is used, and the number of times the prompt word template is modified.

[0232] S53 . For each prompt word template, determine the ratio of the cumulative number of adoptions to the cumulative number of exposures as the adoption rate of all users' historical activity recommendation plans for the target activity theme generated using the prompt word template.

[0233] During specific implementation, for each prompt word template, the activity recommendation module will use the ratio of the cumulative number of adoptions M of historical activity recommendation schemes for the target activity theme by all users in the previous historical time period to the cumulative number of exposures N of historical activity recommendation schemes for the target activity theme corresponding to all users generated using the prompt word template as the adoption rate L of historical activity recommendation schemes for the target activity theme generated using the prompt word template by all users in the previous historical time period.

[0234] Specifically, the adoption rate of historical activity recommendation solutions for the target activity theme generated using the prompt word template by all users in the previous historical time period can be calculated using the following formula:

[0235]

[0236] Wherein, L represents the adoption rate of historical activity recommendation solutions for the target activity theme generated by the prompt word template by all users in the previous historical time period;

[0237] M represents the cumulative number of times all users adopted historical activity recommendations for the target activity theme during the previous historical period;

[0238] N represents the cumulative exposure times of historical activity recommendation plans corresponding to the target activity theme for all users generated using the prompt word template in the previous historical time period.

[0239] S54: Determine the ratio of the number of times the prompt word template is modified to the number of times the prompt word template is used as the modification degree of the prompt word template.

[0240] The modification degree of the prompt word template represents the frequency of modification of the prompt word template.

[0241] During implementation, the activity recommendation module uses, for each prompt word template, the ratio of the number of times the prompt word template has been modified to the number of times the prompt word template has been used in the previous historical time period as the modification degree f of the prompt word template.

[0242] Specifically, the modification degree of the prompt word template can be calculated by the following formula:

[0243]

[0244] Wherein, f represents the modification degree of the prompt word template in the previous historical time period;

[0245] p represents the number of times the prompt word template has been modified in the previous historical time period;

[0246] q represents the number of times the prompt word template is used in the previous historical time period.

[0247] S55. Determine a quality indicator of the prompt word template based on the adoption rate of all users' historical activity recommendation plans for the target activity theme, the modification degree of the prompt word template, and their corresponding weights.

[0248] In specific implementation, the quality index of the prompt word template can be determined by the following formula:

[0249] R(s)=ω1L*ω2f

[0250] Wherein, R(s) represents the quality index of the prompt word template in the previous historical time period;

[0251] L represents the adoption rate of historical activity recommendations for the target activity theme by all users in the previous historical time period;

[0252] ω1 represents the weight of the adoption rate of all users’ historical activity recommendation plans for the target activity theme;

[0253] f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0254] Among them, ω1+ω2=1, ω1 and ω2 can be set arbitrarily, for example, ω1=0.7 and ω2=0.3, which is not limited in the embodiment of the present application.

[0255] S25. Input the target prompt word into the large language model to obtain a target activity recommendation plan corresponding to the target user.

[0256] During specific implementation, the activity recommendation module inputs the generated target prompt words into the large language model to obtain the target activity recommendation plan corresponding to the target user. The large language model is used to plan the activity recommendation plan for the target user based on the target prompt words. The large language model can be obtained by pre-training the activity recommendation module using large-scale text data. The large language model can be, but is not limited to, a ChatGPT (Chat Generative Pre-trained Transformer, generative pre-training transform) model, a GPT-3 model, a GPT-4 model, a BLOOM model or a Deepseek model, etc., and this embodiment of the application does not limit this. In the embodiment of the present application, the large language model can be trained by a large amount of natural language text data. The large language model can understand and generate natural language by learning the language pattern, grammar and semantics of the text data.

[0257] In one embodiment, in order to further improve the usability of activity recommendation plans, historical activity recommendation plans corresponding to target activity themes of multiple other users can be extracted from the activity planning knowledge base and input into the large language model together with the target prompt words as auxiliary information to assist the large language model in generating target activity recommendation plans corresponding to the target user.

[0258] S26: Send the target activity recommendation plan to the display module for display.

[0259] During specific implementation, the activity recommendation module returns the target activity recommendation solution output by the large language model to the presentation module in the client or computing device, and the presentation module presents the target activity recommendation solution.

[0260] The following is an example of super experience activity recommendation in the insurance industry. The client sends a super activity recommendation request to the computing device through the installed super body intelligent assistant application. The display interface of the super body intelligent assistant application is as follows: Figure 6As shown, a conversational page can be used. When a user clicks the Super Body Intelligent Assistant app icon to enter the presentation page, a conversation message automatically triggers: "Hello, I'm Super Body Intelligent Assistant. Feel free to ask any questions." It then displays three frequently asked questions categories. Users can click the drop-down arrow to display questions within each category. These frequently asked questions are configurable. Upon entering the presentation page, two buttons appear: "AI Dialogue" and "Plan an Event." The default is "AI Dialogue." An input box appears below: Enter your chat content to initiate an event recommendation request. Clicking the "Plan an Event" button automatically triggers the conversation message "Plan an Event" and a configuration box: Select "Departure Point" and "Destination." The "Departure Point" can be the automatically located current location. Clicking this allows you to modify the location, triggering a pop-up window for searching cities. Text input is supported. The user selects a departure city. After selecting, exit the departure point and click Cancel to return. If the user's location has not been obtained, a pop-up window will be triggered, asking the user to allow location access. "Destination" directly displays a text input box. Click on the text input to trigger a pop-up window to search for cities. Text input is supported. Select the destination city. Multiple cities can be selected. After selection, the input box displays the tag of the selected city. Each tag can be deleted by clicking on it. Click Cancel to return. Among them, "Departure Place" and "Destination" can also be entered directly by the user. The user clicks "Next" according to the prompt. The "Number of Activity Days" is displayed on the interface. The default is 1 day. You can click the "+" or "-" sign to increase or decrease. The minimum setting is 1 day and the maximum setting is 100 days. The user clicks "Next" and the interface displays "Number of participants, activity budget, activity name, activity type". These four non-mandatory contents have no default values. The first three are input boxes. "Activity Type" is multiple choice. You can choose "Longevity Super Body" and "Healthy Super Body". If you choose Healthy Super Body, the final generated plan must include at least one Healthy Super Body organization.

[0261] The user clicks on the intelligent generation to jump to the loading interface, and sends the above-filled super-body activity recommendation request information to the background computing device. After receiving the above-filled super-body activity recommendation request information, the activity recommendation module in the computing device obtains the prompt word template corresponding to the super-body activity plan from the prompt word template library, obtains the user preference label from the activity planning knowledge base, adds the above-filled information and the user preference label to the corresponding position of the prompt word template, generates the target prompt word, inputs the target prompt word into the large language model, and outputs the target super-body activity recommendation plan. The large language model can adopt the Deepseek model. The activity recommendation module in the computing device returns the generated target super-body activity recommendation plan to the super-body intelligent assistant application of the client. The display interface of the super-body intelligent assistant application displays the activity plan details page, and outputs the target super-body activity recommendation plan, such as Figure 7 shown.

[0262] The display interface of the Super Body Intelligent Assistant application can also initiate activity recommendation requests in the form of multi-round dialogues, such as Figure 8 As shown, if the user enters text containing the words "planning an event," a request for a super-body activity recommendation is sent to the computing device. The computing device determines whether there is information such as the departure and destination (the destination can be a city or a super-body organization), the number of days for the event, the number of participants, the budget for the event, and the type of event. If such information is available, the computing device will respond with "OK, we received your request. We are generating an activity plan for you. Click the button below to view it." The user clicks "View Plan" and is redirected to the loading screen. If any parameter is missing, the activity recommendation module in the computing device can add additional questions, which correspond to the following questions:

[0263] "Where do you plan to depart from?"

[0264] "Which cities or super-body scenes are the destinations?"

[0265] “How many days does the event last?”

[0266] “How many people are there at the event?”

[0267] “What’s the budget for the event?”

[0268] "Do we need a long-lived superbody or a healthy superbody?"

[0269] If the user enters "Unclear," "Don't Know," or "Undecided," the system will respond with "Okay, we've received your request. We're generating an activity plan for you. Click the button below to view it." The confirmed information and user preference tags will be added to the prompt word template to obtain the target prompt word. This target prompt word will be input into the large language model to generate the target Super Body activity recommendation plan. If the user does not provide any information, a pre-set default Super Body activity recommendation plan will be returned to the client.

[0270] If the user is not satisfied with the generated super-body activity recommendation, he or she can also enter feedback in the current display interface, such as Figure 9 As shown, you can click the "dislike" button to trigger a pop-up window, prompting you to give feedback if you are not satisfied with the solution. The user enters the feedback and saves it to the system. If the user wants to regenerate a new recommendation solution, click Figure 9 The "Regenerate" button in triggers the computing device to regenerate a new super-body activity recommendation plan.

[0271] An activity recommendation method provided by an embodiment of the present application comprises an activity recommendation module receiving activity recommendation request information, the activity recommendation request information including target user identification information, target activity theme information and activity constraint information, obtaining a prompt word template corresponding to the target activity theme from a prompt word template library, obtaining a target user preference label corresponding to the target user identification information from an activity planning knowledge base, the target user preference label being determined based on the target user's feedback index on historical activity recommendation data, the feedback index representing the target user's interest in the historical activity recommendation data, generating a target prompt word based on the target activity theme information, activity constraint information, the target user preference label and the prompt word template, inputting the target prompt word into a large language model, obtaining a target activity recommendation plan corresponding to the target user, the large language model being used to plan an activity recommendation plan for the target user based on the target prompt word, and sending the target activity recommendation plan to a presentation module for presentation. In an embodiment of the present application, corresponding prompt word templates are set in advance for different activity themes, and user feedback indicators for historical activity recommendation data are recorded in the activity planning knowledge base. User preference labels are determined based on the user feedback indicators for historical activity recommendation data. When the activity recommendation module receives an activity recommendation request, the prompt word template corresponding to the target activity theme is obtained from the prompt word template library, and the target user preference label, target activity theme information, and activity constraint information obtained from the activity planning knowledge base are combined to generate target prompt words. The target recommended activity plan corresponding to the target user is predicted based on the target prompt words and the large language model, and is displayed to the user through the display module. Since the generated target prompt words are more in line with the preferences of the target user, while improving the efficiency of activity plan planning, the planned activity plan can be recommended to the target user more accurately, thereby enhancing the user's experience and feelings about the activity.

[0272] Based on the same inventive concept, an embodiment of the present application further provides an activity recommendation device. Since the principle of solving the problem by the above-mentioned activity recommendation device is similar to that of the above-mentioned activity recommendation method, the implementation of the above-mentioned device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0273] like Figure 10 As shown, it is a structural diagram of the activity recommendation device provided by the embodiment of the present application, and the activity recommendation device can be applied to Figure 1 In the activity recommendation module of the activity recommendation system shown, the device may include:

[0274] A receiving module 61 is configured to receive activity recommendation request information, wherein the activity recommendation request information includes target user identification information, target activity theme information, and activity constraint information;

[0275] A first acquisition module 62 is configured to acquire a prompt word template corresponding to the target activity theme from a prompt word template library;

[0276] A second acquisition module 63 is configured to acquire a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined based on a feedback index of the target user on the historical activity recommendation data, wherein the feedback index represents the target user's interest in the historical activity recommendation data;

[0277] A generating module 64 is configured to generate a target prompt word according to the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template;

[0278] A recommendation module 65 is configured to input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, wherein the large language model is used to plan an activity recommendation plan for the target user based on the target prompt word;

[0279] The sending module 66 is used to send the target activity recommendation plan to the display module for display.

[0280] In one embodiment, the second acquisition module is specifically used to determine the target user's feedback index on the historical activity recommendation data in the following manner: obtaining the historical activity recommendation plan of the target activity theme corresponding to the target user from the activity planning knowledge base; counting the number of exposures of the historical activity recommendation plan of the target activity theme corresponding to the target user and the number of adoptions of the historical activity recommendation plan of the target activity theme by the target user; obtaining the browsing time of the historical activity recommendation plan of the target activity theme by the target user; and determining the target user's feedback index on the historical activity recommendation data based on the number of exposures, the number of adoptions and the browsing time.

[0281] In one embodiment, the second acquisition module 63 is specifically used to determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the target user's historical activity recommendation schemes for the target activity theme; and determine the target user's feedback index for the historical activity recommendation data based on the target user's adoption rate of the historical activity recommendation schemes for the target activity theme and the browsing time.

[0282] In one embodiment, the second acquisition module 63 is specifically configured to calculate the target user's feedback index for the historical activity recommendation data using the following formula:

[0283] W=l*logC

[0284] Wherein, W represents the target user's feedback indicator on the historical activity recommendation data;

[0285] l represents the adoption rate of the target user’s historical activity recommendation solutions for the target activity theme, m represents the number of times the target user adopts the historical activity recommendation scheme of the target activity theme, and n represents the number of times the target user is exposed to the historical activity recommendation scheme of the target activity theme;

[0286] C represents the browsing time of the target user for the historical activity recommendation solutions of the target activity theme.

[0287] In one embodiment, the second acquisition module 63 is specifically used to determine the target user preference label based on the target user's feedback index on the historical activity recommendation data in the following manner: if it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, then extracting the first target activity object from the historical activity recommendation scheme adopted by the target user as the target user's interest label; if it is determined that the target user's feedback index on the historical activity recommendation data is less than the preset threshold, then extracting the second target activity object from the historical activity recommendation scheme not adopted by the target user as the target user's non-interest label.

[0288] In one embodiment, the device further comprises:

[0289] a determination module configured to, if it is determined that the target activity theme includes multiple prompt word templates, determine a quality indicator of each prompt word template based on historical activity recommendation data of the target activity theme for all users in a previous historical time period;

[0290] The selection module is used to select the prompt word template with the highest quality index.

[0291] In one embodiment, the determination module is specifically configured to obtain, from the activity planning knowledge base, historical activity recommendation schemes for the target activity theme corresponding to all users within a previous historical time period; respectively count the cumulative number of exposures of the historical activity recommendation schemes for the target activity theme corresponding to each prompt word template for the target activity theme generated by all users within the previous historical time period, the cumulative number of adoptions of the historical activity recommendation schemes for the target activity theme by all users, the number of uses of the prompt word template, and the number of modifications of the prompt word template; for each prompt word template, determine, by the ratio of the cumulative number of adoptions to the cumulative number of exposures, an adoption rate of the historical activity recommendation schemes for the target activity theme generated by the prompt word template by all users; determine, by the ratio of the number of modifications of the prompt word template to the number of uses of the prompt word template, a modification degree of the prompt word template, the modification degree of the prompt word template representing the frequency of modifications to the prompt word template; and determine a quality indicator of the prompt word template based on the adoption rate of the historical activity recommendation schemes for the target activity theme by all users, the modification degree of the prompt word template, and their corresponding weights.

[0292] In one embodiment, the determining module is specifically configured to determine the quality index of the prompt word template using the following formula:

[0293] R(s)=ω1L*ω2f

[0294] Wherein, R(s) represents the quality index of the prompt word template;

[0295] L represents the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, and ω1 represents the weight of the adoption rate of the historical activity recommendation schemes of the target activity theme by all users;

[0296] f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

[0297] Based on the same technical concept, the embodiment of the present application further provides an electronic device 700, referring to Figure 11 As shown, electronic device 700 is used to implement the activity recommendation method described in the above method embodiments. Electronic device 700 of this embodiment may include: memory 701, processor 702, and a computer program stored in the memory and executable on the processor, such as an activity recommendation program. When the processor executes the computer program, the steps of each of the above activity recommendation method embodiments are implemented.

[0298] The specific connection medium between the memory 701 and the processor 702 is not limited in the embodiment of the present application. Figure 11 In the embodiment, the memory 701 and the processor 702 are connected via a bus 703. The bus 703 is connected to the processor 702 via a bus 703. Figure 11 The bus 703 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 11 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0299] Memory 701 may be a volatile memory, such as random-access memory (RAM); a non-volatile memory, such as read-only memory, flash memory, a hard disk drive (HDD), or a solid-state drive (SSD); or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 701 may be a combination of the above memories.

[0300] The processor 702 is configured to implement the activity recommendation method provided in the embodiment of the present application.

[0301] An embodiment of the present application also provides a computer-readable storage medium that stores computer-executable instructions required to execute the above-mentioned processor, which includes a program required to execute the above-mentioned processor.

[0302] In some possible implementations, various aspects of the activity recommendation method provided in the present application may also be implemented in the form of a program product, which includes program code. When the program product is run on an electronic device, the program code is used to enable the electronic device to execute the steps of the activity recommendation method according to various exemplary implementations of the present application described above in this specification.

[0303] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, devices, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0304] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (apparatus), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0305] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0306] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0307] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0308] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. An activity recommendation system, characterized in that: The system includes a prompt word template library, an activity planning knowledge base, an activity recommendation module and a display module; wherein: The prompt word template library is used to store prompt word templates corresponding to preset activity themes; The activity planning knowledge base is used to store the correspondence between user identification information and the user's historical activity recommendation data and user preference tags; The activity recommendation module is configured to receive activity recommendation request information, the activity recommendation request information including target user identification information, target activity theme information, and activity constraint information; obtain a prompt word template corresponding to the target activity theme from a prompt word template library; obtain a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, the target user preference tag being determined based on a feedback indicator of the target user on historical activity recommendation data, the feedback indicator representing the target user's interest in the historical activity recommendation data; generate a target prompt word based on the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template; input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, the large language model being used to plan an activity recommendation plan for the target user based on the target prompt word; The display module is used to display the target activity recommendation plan.

2. The system according to claim 1, wherein The activity recommendation module is specifically configured to determine the target user's feedback index for the historical activity recommendation data by: obtaining the historical activity recommendation plan for the target activity theme corresponding to the target user from the activity planning knowledge base; Counting the number of exposures of historical activity recommendation schemes of the target activity theme corresponding to the target user and the number of adoptions of historical activity recommendation schemes of the target activity theme by the target user; Obtaining the target user's browsing time for historical activity recommendations for the target activity theme; A feedback index of the target user for the historical activity recommendation data is determined according to the number of exposures, the number of adoptions, and the browsing time.

3. The system according to claim 2, wherein: The activity recommendation module is specifically configured to determine the ratio of the number of adoptions to the number of exposures as the adoption rate of the target user's historical activity recommendation solutions for the target activity theme; A feedback indicator of the target user on the historical activity recommendation data is determined according to the adoption rate of the target user's historical activity recommendation solutions for the target activity theme and the browsing time.

4. The system according to claim 3, wherein: The activity recommendation module is specifically configured to calculate the target user's feedback index on the historical activity recommendation data using the following formula: W=l*logC Wherein, W represents the target user's feedback indicator on the historical activity recommendation data; l represents the adoption rate of the target user’s historical activity recommendation solutions for the target activity theme, m represents the number of times the target user adopts the historical activity recommendation scheme of the target activity theme, and n represents the number of times the target user is exposed to the historical activity recommendation scheme of the target activity theme; C represents the browsing time of the target user for the historical activity recommendation solutions of the target activity theme.

5. The system according to any one of claims 2 to 4, characterized in that: The activity recommendation module is specifically configured to determine the target user preference tag based on the target user's feedback indicator on historical activity recommendation data in the following manner: If it is determined that the target user's feedback index on the historical activity recommendation data is greater than or equal to a preset threshold, extracting a first target activity object from the historical activity recommendation scheme adopted by the target user as a target user interest tag; If it is determined that the target user's feedback index on the historical activity recommendation data is less than the preset threshold, a second target activity object is extracted from the historical activity recommendation schemes not adopted by the target user as a target user not interested label.

6. The system according to claim 1, wherein: The activity recommendation module is further configured to, if it is determined that the target activity theme includes multiple prompt word templates, determine a quality indicator of each prompt word template based on historical activity recommendation data of the target activity theme for all users in the previous historical time period; Select the prompt word template with the highest quality index.

7. The system according to claim 6, wherein: The activity recommendation module is specifically configured to obtain, from the activity planning knowledge base, historical activity recommendation plans for the target activity theme corresponding to all users in the previous historical time period; and to count the cumulative number of exposures of the historical activity recommendation plans for the target activity theme corresponding to all users generated using each prompt word template corresponding to the target activity theme in the previous historical time period, the cumulative number of adoptions of the historical activity recommendation plans for the target activity theme by all users, the number of uses of the prompt word template, and the number of modifications of the prompt word template. For each prompt word template, the ratio of the cumulative number of adoptions to the cumulative number of exposures is used to determine the adoption rate of the historical activity recommendations for the target activity theme generated by all users using the prompt word template. The ratio of the number of modifications to the prompt word template to the number of uses of the prompt word template is used to determine the modification degree of the prompt word template. The modification degree of the prompt word template represents the frequency of modification to the prompt word template. The quality index of the prompt word template is determined according to the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, the modification degree of the prompt word template, and the corresponding weights.

8. The system according to claim 7, wherein: The activity recommendation module is specifically configured to determine the quality index of the prompt word template using the following formula: R(s)=ω1L*ω2f Wherein, R(s) represents the quality index of the prompt word template; L represents the adoption rate of the historical activity recommendation schemes of the target activity theme by all users, and ω1 represents the weight of the adoption rate of the historical activity recommendation schemes of the target activity theme by all users; f represents the modification degree of the prompt word template, and ω2 represents the weight of the modification degree of the prompt word template.

9. An activity recommendation method, characterized in that: include: receiving an activity recommendation request message, wherein the activity recommendation request message includes target user identification information, target activity theme information, and activity constraint information; Obtaining a prompt word template corresponding to the target activity theme from a prompt word template library; Obtaining a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined based on a feedback indicator of the target user on the historical activity recommendation data, wherein the feedback indicator represents the target user's interest in the historical activity recommendation data; generating a target prompt word according to the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template; Inputting the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, wherein the large language model is used to plan an activity recommendation plan for the target user based on the target prompt word; The target activity recommendation plan is sent to the display module for display.

10. An activity recommendation device, characterized in that: include: A receiving module, configured to receive activity recommendation request information, wherein the activity recommendation request information includes target user identification information, target activity theme information, and activity constraint information; A first acquisition module is used to acquire a prompt word template corresponding to the target activity theme from a prompt word template library; A second acquisition module is configured to acquire a target user preference tag corresponding to the target user identification information from an activity planning knowledge base, wherein the target user preference tag is determined based on a feedback index of the target user on the historical activity recommendation data, wherein the feedback index represents the target user's interest in the historical activity recommendation data; A generating module, configured to generate a target prompt word according to the target activity theme information, the activity constraint information, the target user preference tag, and the prompt word template; a recommendation module, configured to input the target prompt word into a large language model to obtain a target activity recommendation plan corresponding to the target user, wherein the large language model is used to plan an activity recommendation plan for the target user based on the target prompt word; The sending module is used to send the target activity recommendation plan to the display module for display.

Images