Object recommendation method, object recommendation device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides an object recommendation method, an object recommendation device, electronic equipment and a storage medium, and belongs to the technical field of financial science and technology. The method comprises the steps of obtaining content interaction data obtained by interaction of a target object on target promotion content in an application; obtaining a content interaction object of the target object based on the content interaction data; performing object category detection on the content interaction object based on the content interaction sub-data to obtain an interaction object category; obtaining candidate object features of the content interaction object and historical interaction data of the content interaction object in the application; determining a preference matching score of the content interaction object and the target promotion content based on the historical interaction data; performing object conversion prediction based on the interactive object category, the candidate object features and the preference matching score to obtain a target conversion score of the content interactive object; and performing object recommendation on the target object based on the target conversion score. According to the embodiment of the invention, the accuracy and efficiency of object recommendation can be improved.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and in particular to an object recommendation method and object recommendation device, electronic device, and storage medium. Background Art

[0002] Object recommendation refers to the process of recommending objects that may be willing to establish associations for the target object, thereby increasing the association rate between objects in the application. For example, in the scenario of adding sales personnel in financial technology, the recommended objects can be objects that may become sales personnel. For another example, in the scenario of adding friends, the recommended objects can be objects that are willing to become friends with the target object.

[0003] At present, in the scenario of recommending objects for specific occupations, objects that are closely related to the target object (i.e., objects belonging to a specific occupation) can usually be used as candidates, and recommendations can be made based on the results of interviews with the candidates. Alternatively, candidates that may be of interest can be selected through the Internet, and these candidates can be sorted by a single standard to be recommended and displayed to the target object. However, the first method is prone to low attendance or participation of candidates, which reduces the efficiency of object recommendation. In addition, due to specific occupational requirements, multiple factors need to be considered when selecting objects. The second method has poor accuracy in object recommendation for this scenario and cannot effectively promote the conversion rate of recommendation processing (i.e., the probability that the target object establishes an effective association with the recommended object after being recommended to the target object). Therefore, how to improve the accuracy and efficiency of object recommendation has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose an object recommendation method and an object recommendation device, an electronic device, and a storage medium, aiming to improve the accuracy and efficiency of object recommendation.

[0005] To achieve the above objective, a first aspect of an embodiment of the present application proposes an object recommendation method, the method comprising:

[0006] Acquire content interaction data obtained by a target object interacting with target promotion content in an application, where the target promotion content includes a preset promotion category;

[0007] Acquire a content interaction object of the target object based on the content interaction data, wherein the content interaction data includes content interaction sub-data of the content interaction object;

[0008] Performing object category detection on the content interaction object based on the content interaction sub-data to obtain an interaction object category, wherein the interaction object category is used to indicate an emotional direction of the interaction between the content interaction object and the target promotion content in the application;

[0009] Acquire candidate object features of the content interactive object and historical interaction data of the content interactive object in the application;

[0010] Determine a preference matching score between the content interaction object and the target promotion content based on the historical interaction data;

[0011] Perform object conversion prediction based on the interactive object category, the candidate object features and the preference matching score to obtain a target conversion score for the content interactive object, wherein the target conversion score is used to characterize the degree to which the content interactive object is converted into an object matching the preset promotion category;

[0012] An object is recommended to the target object based on the target conversion score.

[0013] In some embodiments, the recommending an object to the target object based on the target conversion score includes:

[0014] Performing grade matching based on the target conversion score and the preset score level to obtain the recommended object level of the content interaction object;

[0015] Sort the content interaction objects based on the recommendation object levels and the target conversion scores to obtain a recommendation object sequence;

[0016] A target recommended object is selected from the recommended object sequence based on a preset recommendation quantity, and the target recommended object is recommended to the target object.

[0017] In some embodiments, determining the preference matching score between the content interaction object and the target promotion content based on the historical interaction data includes:

[0018] Acquire the associated object data of the content interaction object from the historical interaction data;

[0019] Determine the total number of associated objects, the total number of objects with the same attributes, and the total number of commonly associated objects of the content interactive object based on the associated object data, the total number of associated objects being used to represent the number of objects associated with the content interactive object in the application, the total number of objects with the same attributes being used to represent the number of objects with the same preset promotion category associated with the content interactive object in the application, and the total number of commonly associated objects being used to represent the number of objects commonly associated between the content interactive object and the target object in the application;

[0020] Determine the interaction frequency and main interaction type between the content interaction object and the target object in the application based on the content interaction sub-data;

[0021] Based on the total number of associated objects, the total number of objects with the same attribute, the number of commonly associated objects, the interaction frequency and the main interaction type, the preference matching score between the content interaction object and the target promotion content is determined.

[0022] In some embodiments, determining the preference matching score between the content interaction object and the target promotion content based on the total number of associated objects, the total number of objects with the same attribute, the number of commonly associated objects, the interaction frequency, and the main interaction type includes:

[0023] Determining a first matching score of the content interaction object based on the total number of associated objects;

[0024] Determining a second matching score of the content interaction object based on the total number of objects with the same attribute;

[0025] Determining a third matching score of the content interaction object based on the number of commonly associated objects;

[0026] Based on the interaction frequency, determining a fourth matching score of the content interaction object;

[0027] Based on the primary interaction type, determining a fifth matching score of the content interaction object;

[0028] The preference matching score between the content interactive object and the target promotion content is determined based on the first matching score, the second matching score, the third matching score, the fourth matching score and the fifth matching score of the content interactive object.

[0029] In some embodiments, before obtaining the content interaction data obtained by the target object interacting with the target promotion content in the application, the method further includes:

[0030] Get the object type of the target object and generate content conditional text;

[0031] Performing content generation on the object type and the generated content condition text based on a preset content generation model to obtain candidate promotion content, wherein the candidate promotion content is stored in a content library of the content generation model;

[0032] Select target promotion content from the candidate promotion content, and mark the target promotion content in the content library to update the priority of the target promotion content in the content library.

[0033] In some embodiments, performing object category detection on the content interactive object based on the content interactive sub-data to obtain the interactive object category includes:

[0034] Acquire content interaction text from the content interaction sub-data;

[0035] Performing text segmentation on the content interaction text to obtain content interaction words;

[0036] Performing word vectorization on the content interaction words to obtain interaction word vectors;

[0037] Get the preset word vector of the preset category keyword;

[0038] Extract key features of the interactive word vector to obtain an interactive keyword vector;

[0039] Performing vector similarity calculation on the preset word vector and the interactive keyword vector to obtain word vector similarity;

[0040] The interactive object category is determined based on the word vector similarity.

[0041] In some embodiments, performing object conversion prediction based on the interactive object category, the candidate object features, and the preference matching score to obtain a target conversion score of the content interactive object includes:

[0042] Determine interaction strategy data for the content interaction object based on the interaction object category and the content interaction sub-data;

[0043] Obtain object feedback data of the content interaction object to the interaction strategy data;

[0044] The target conversion score of the content interactive object is determined based on the interactive object category, the candidate object characteristics, the preference matching score and the object feedback data.

[0045] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides an object recommendation device, the device comprising:

[0046] A first acquisition module is used to acquire content interaction data obtained by a target object interacting with a target promotion content in an application, wherein the target promotion content includes a preset promotion category;

[0047] An object acquisition module, configured to acquire a content interaction object of the target object based on the content interaction data, wherein the content interaction data includes content interaction sub-data of the content interaction object;

[0048] A category detection module, used to perform object category detection on the content interaction object based on the content interaction sub-data to obtain an interactive object category, wherein the interactive object category is used to indicate an emotional direction of the interaction between the content interaction object and the target promotion content in the application;

[0049] A second acquisition module, used to acquire candidate object features of the content interactive object and historical interaction data of the content interactive object in the application;

[0050] A score determination module, used to determine a preference matching score between the content interaction object and the target promotion content based on the historical interaction data;

[0051] An object conversion module, configured to predict object conversion based on the interactive object category, the candidate object features and the preference matching score, and obtain a target conversion score for the content interactive object, wherein the target conversion score is used to characterize the degree to which the content interactive object is converted into an object matching the preset promotion category;

[0052] A recommendation module is used to recommend an object to the target object based on the target conversion score.

[0053] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor implements an object recommendation method described in the first aspect when executing the computer program.

[0054] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the object recommendation method described in the first aspect is implemented.

[0055] The present application proposes an object recommendation method and an object recommendation device, an electronic device, and a storage medium, which first obtain content interaction data obtained by a target object interacting with target promotion content in an application, wherein the target promotion content includes a preset promotion category; further, based on the content interaction data, a content interaction object of the target object is obtained, wherein the content interaction data includes content interaction sub-data of the content interaction object; based on the content interaction sub-data, an object category detection is performed on the content interaction object to obtain an interaction object category, wherein the interaction object category is used to indicate the emotional direction of the interaction between the content interaction object and the target promotion content in the application; further, candidate object features of the content interaction object and historical interaction data of the content interaction object in the application are obtained; based on the historical interaction data, a preference matching score between the content interaction object and the target promotion content is determined; further, based on the interaction object category, the candidate object features, and the preference matching score, an object conversion prediction is performed to obtain a target conversion score of the content interaction object, wherein the target conversion score is used to characterize the degree to which the content interaction object is converted into an object matching the preset promotion category; finally, an object recommendation is performed to the target object based on the target conversion score. After determining the content interaction object of the target object, the embodiment of the present application can first perform object category detection on the content interaction object, and then determine the target conversion score of the content interaction object based on the detected category and the determined preference matching score between the content interaction object and the target promotion content. Compared with the related art that does not consider various factors related to the preset promotion category when selecting objects, the present application can fully consider various factors related to the target promotion content and the content interaction object itself by determining the preference matching score and the target conversion score, thereby improving the accuracy and efficiency of object recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is the first flow chart of the object recommendation method provided by the embodiment of the present application;

[0057] Figure 2 is a second flow chart of the object recommendation method provided in an embodiment of the present application;

[0058] Figure 3 yes Figure 1 A flowchart of step S130 in FIG.

[0059] Figure 4 yes Figure 1 A flowchart of step S150 in FIG.

[0060] Figure 5 yes Figure 4 A flowchart of step S440 in FIG.

[0061] Figure 6 yes Figure 1 A flowchart of step S160 in FIG.

[0062] Figure 7 yes Figure 1 A flowchart of step S170 in FIG.

[0063] Figure 8 is a structural diagram of an object recommendation device provided in an embodiment of the present application;

[0064] Fig. 9 It is a hardware structure diagram of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0067] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0068] First, some nouns involved in this application are analyzed:

[0069] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0070] Robotic Process Automation (RPA) is a technology that uses software robots or artificial intelligence workers to automate repetitive tasks. It is designed to automate repetitive, regular and predictable business processes or tasks.

[0071] Chat Generalized Language Model (ChatGLM): Based on the Transformer architecture, ChatGLM learns language patterns through an autoregressive model and can generate coherent and natural responses based on context. ChatGLM can be pre-trained with a large amount of text data to learn language patterns and generate text.

[0072] Natural Language Processing (NLP) is an important branch of artificial intelligence, which aims to enable computers to understand and process human language. NLP technology achieves natural interaction with humans by analyzing, understanding and generating human language.

[0073] Object recommendation refers to the process of recommending objects that may be willing to establish associations for the target object, thereby increasing the association rate between objects in the application. For example, in the scenario of adding sales personnel in financial technology, the recommended objects can be objects that may become sales personnel. For another example, in the scenario of adding friends, the recommended objects can be objects that are willing to become friends with the target object.

[0074] At present, in the scenario of recommending objects for specific occupations, objects that are closely related to the target object (i.e., objects belonging to a specific occupation) can usually be used as candidates, and recommendations can be made based on the results of interviews with the candidates. Alternatively, candidates that may be of interest can be selected through the Internet, and these candidates can be sorted by a single standard to be recommended and displayed to the target object. However, the first method is prone to low attendance or participation of candidates, which reduces the efficiency of object recommendation. In addition, due to specific occupational requirements, multiple factors need to be considered when selecting objects. The second method has poor accuracy in object recommendation for this scenario and cannot effectively promote the conversion rate of recommendation processing (i.e., the probability that the target object establishes an effective association with the recommended object after being recommended to the target object). Therefore, how to improve the accuracy and efficiency of object recommendation has become a technical problem that needs to be solved urgently.

[0075] Based on this, the embodiments of the present application provide an object recommendation method and an object recommendation device, an electronic device, and a storage medium, which can improve the accuracy and efficiency of object recommendation.

[0076] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0077] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0078] The object recommendation method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The object recommendation method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or can be configured as a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN) and big data and artificial intelligence platforms and other basic cloud computing services; the software can be an application that implements the object recommendation method, etc., but is not limited to the above forms.

[0079] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0080] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to the identity or characteristics of the object, such as object attribute information, object identity information, object association information, etc., the permission or consent of the object will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. In addition, when the embodiment of the present application needs to obtain sensitive personal information of the object, the separate permission or separate consent of the object will be obtained through a pop-up window or jump to a confirmation page. After clearly obtaining the separate permission or separate consent of the object, the necessary object-related data for enabling the normal operation of the embodiment of the present application will be obtained.

[0081] See also Figure 1 , Figure 1 is an optional flow chart of the object recommendation method provided in the embodiment of the present application. In some embodiments, Figure 1 The method may include but is not limited to steps S110 to S170:

[0082] Step S110, obtaining content interaction data obtained by the target object interacting with the target promotion content in the application;

[0083] Step S120, acquiring a content interaction object of the target object based on the content interaction data;

[0084] Step S130, performing object category detection on the content interaction object based on the content interaction sub-data to obtain the interaction object category;

[0085] Step S140, obtaining candidate object features of the content interactive object and historical interaction data of the content interactive object in the application;

[0086] Step S150, determining a preference matching score between the content interaction object and the target promotion content based on the historical interaction data;

[0087] Step S160, performing object conversion prediction based on the interactive object category, candidate object features and preference matching score to obtain a target conversion score for the content interactive object;

[0088] Step S170: recommending an object to a target object based on the target conversion score.

[0089] In step S110 of some embodiments, the target object refers to the object for which the recommended object needs to be obtained. Application refers to software that can upload target promotion content, and the target promotion content can be pictures, videos, voices, etc., and its form is not specifically limited. For example, in the intelligent sales recruitment scenario of financial technology, in order to attract more objects who want to be called sales personnel, the target object can upload the obtained target promotion content to the application, and the target promotion content can be a motivational video related to the sales industry or entrepreneurship. In this way, other objects using the application can view the target promotion content, and can interact with the target promotion content (such as likes, comments, forwarding, etc.) to achieve interaction with the target object. Content interaction data refers to data generated by other objects using the application to interact with the target promotion content, such as specific comments on the target promotion content, whether to like, whether to forward, the number of forwardings, etc.

[0090] It should be noted that, since the target object may upload more than one target promotional content in the application, the content interaction data at this time may refer to the data obtained after counting the interactions of all target promotional content uploaded by the target object, for example, all comments, all likes, all likes, all forwarded content, all forwarding times, etc. of the target promotional content by other objects using the application.

[0091] It should be noted that the target promotion content includes a preset promotion category, which can be used to indicate the key attribute characteristics of the target promotion content. For example, the target promotion content is mainly for selling items or attracting talents, etc., so that other objects viewing the target promotion content can quickly understand the promotion needs and goals of the content.

[0092] It should be noted that in order to allow more objects to view the target promotion content uploaded by the target object, this application can also automatically interact with the content uploaded by the objects associated with the target object according to the preset strategy of the target object through RPA, such as forwarding and commenting, so as to obtain more other objects to view the target promotion content of the target object and improve the efficiency of object recommendation. For example, in the intelligent sales recruitment scenario of financial technology, the target object follows two objects in the application account, namely account object Z1 and account object Z2. This application can set RPA execution strategies through RPA, such as forwarding videos within 1 hour after the new video is released, leaving messages on the first 5 newly uploaded videos, etc. At this time, if account object Z1 uploads a video V1, the target object can automatically forward and leave messages within 1 hour after the video V1 is released.

[0093] See also Figure 2 , Figure 2is another optional flow chart of the object recommendation method provided by the embodiment of the present application. In some embodiments, before step S110, the object recommendation method provided by the embodiment of the present application may further include steps S210 to S230:

[0094] Step S210, obtaining the object type of the target object and generating content condition text;

[0095] Step S220, performing content generation on the object type and the generated content condition text based on a preset content generation model to obtain candidate promotion content;

[0096] Step S230 , selecting target promotion content from the candidate promotion content, and marking the target promotion content in the content library to update the priority of the target promotion content in the content library.

[0097] In step S210 of some embodiments, the object type is used to characterize the object attribute characteristics of the target object or the content attribute characteristics of the target object demand promotion. For example, the object attribute characteristics may include the professional attribute characteristics of the object (such as salesperson, engineer, etc.), and the object content characteristics may characterize the key characteristics of the content used for promotion (such as entrepreneurship, recruitment, etc.). In this way, it can help the model understand how to generate content effectively. Generating content conditional text refers to related text condition information or demand information that affects content generation, for example, "Please generate entrepreneurial words for me, and the number of words should be controlled within 30 words."

[0098] In step S220 of some embodiments, the present application may combine a pre-trained content generation model to generate content, and the content generation model may be constructed based on the model structure of ChatGLM, or based on model structures such as machine learning models and natural language generation models, without specific limitation. Specifically, the object type and the generated content condition text may be input into the content generation model, and content generation may be performed on the object type and the generated content condition text based on the content generation model. These contents may cover different styles, tones, and information to ensure that there is diversity to choose from. The generated candidate promotion content will be stored in the content library of the content generation model for subsequent use. Among them, the content library also stores historically generated promotion content to give the target object more content choices. For example, in the scenario of intelligent recruitment of financial technology, in order to stimulate the potential willingness of candidates by showing the advantages of sales personnel's career development, the target object at this time can be a sales person or a human resources object of the enterprise. Then the prompt of the object type input by the target object can be "I am an insurance agent", and the generated content condition text can be "I am preparing to generate a video on the theme of entrepreneurship and inspiration, hoping to attract comments and attention from interested netizens. Please generate corresponding forwarding words for me, and the number of words should be controlled within 30 words." Further, the content generation model can generate at least one candidate promotion content based on the input object type and the generated content condition text.

[0099] In step S230 of some embodiments, further, one or more most suitable contents may be selected from the candidate promotion contents as the target promotion contents. The candidate promotion contents may include content generated based on the current object type and the generated content condition text, and may also include historical content already stored in the content library. Furthermore, the system used by the object recommendation method of the present application may perform content tagging on the target promotion contents in the content library, record the selection results of the target objects, and update the priority of the target promotion contents in the content library, so that they can be weighted in the next output content sorting.

[0100] It should be noted that the starting score of the candidate promotion content stored in the content library is 0. Each selection by the target object will add 1 point to the selected promotion content. Next time, the top n (n is a positive integer) promotion content will be displayed first from high to low according to the score, and the newly generated candidate promotion content will also appear in the ranking. Among them, since the content generated by the model may be the same as the historical promotion content stored in the content library, after obtaining the content ranking list, the content in the list can be deduplicated and the top ranking one can be retained to update the content ranking list and improve the efficiency of content selection.

[0101] It should be noted that each selection result of the target object can update the score weight of each candidate promotion content, so that different priorities will be given in the next display. For the content generation model, the target object's preference for content is not known at the beginning, so the content generation model can be continuously updated according to the marked content library to improve the matching degree between the generated content and the needs of the target object, and the order of displaying the generated content can be continuously adjusted to achieve the generated content that better meets the preferences of the target object.

[0102] It should be noted that the selection criteria for selecting one or more most suitable contents from the generated candidate promotional contents may include the relevance, attractiveness, expected effect, etc. of the contents, without specific limitation.

[0103] It should be noted that in addition to marking the target promotion content as a whole, the present application can also extract key content information from the target promotion content and mark the key content information to store the marked key content information in the content library, so that the key content information can be combined according to the input object type and the generated content conditional text to improve the accuracy of content generation.

[0104] In the above embodiment, the present application can generate content based on object type and generated content conditional text, and select more matching target promotion content in combination with the content and content tags already stored in the content library, which can improve the accuracy of content generation and content selection efficiency, thereby improving the accuracy of object recommendation.

[0105] It should be noted that, in actual application, the present application can set the content generation model in a functional module of the application. The target object can trigger the generation control (such as "start generation") corresponding to the functional module by inputting a prompt and a content generation requirement in the prompt input box corresponding to the functional module, thereby triggering the subsequent workflow and generating a content ranking list containing multiple candidate promotional contents for the target object to select.

[0106] In step S120 of some embodiments, the content interaction data may include content interaction sub-data of multiple content interaction objects, and the content interaction sub-data refers to the data generated after each content interaction object interacts with the target promotion content in the application. For example, for the target promotion content uploaded by the target object in the application, there are 10 objects commenting on the target promotion content, 3 objects forwarding the target promotion content, and 4 objects liking the target promotion content, and there are no duplicate objects among the objects that comment, forward and like. In this way, the content interaction data at this time may include content interaction sub-data of 17 objects.

[0107] It should be noted that if the same content interaction object both comments on and forwards the target promotion content, the content interaction sub-data corresponding to the content interaction object includes both the comment interaction data and the forwarding interaction data.

[0108] In step S130 of some embodiments, the interactive object category is used to indicate the emotional direction of the interaction between the content interactive object and the target promotion content in the application. The interactive object category may include a positive interaction category and a negative interaction category. The positive interaction category is used to characterize that the emotional direction of the content interactive object to the target promotion content is positive, that is, expressing interest, recognition or positive emotions, for example, likes of the content, comments including positive words such as "great", "strongly agree", "consultation activity details", etc. The negative interaction category is used to characterize that the emotional direction of the content interactive object to the target promotion content is negative, that is, expressing disinterest, disapproval or negative emotions, such as comments including negative words such as "dislike" and "disagree".

[0109] See also Figure 3 , Figure 3 It is an optional flow chart of step S130 provided in an embodiment of the present application. In some embodiments, step S130 may specifically include steps S310 to S370:

[0110] Step S310, obtaining content interaction text from content interaction sub-data;

[0111] Step S320, performing text segmentation on the content interaction text to obtain content interaction words;

[0112] Step S330, performing word vectorization on the content interaction words to obtain interaction word vectors;

[0113] Step S340, obtaining a preset word vector of a preset category keyword;

[0114] Step S350, extracting key features from the interactive word vector to obtain an interactive keyword vector;

[0115] Step S360, calculating the vector similarity between the preset word vector and the interactive keyword vector to obtain the word vector similarity;

[0116] Step S370, determining the interactive object category based on the word vector similarity.

[0117] In step S310 of some embodiments, when determining the object interaction category, it can be determined based on the content interaction text (i.e., the comments on the content). If the content interaction sub-data does not contain the content interaction text, the content operation information can be obtained from the content interaction sub-data. At this time, the content operation information can be forwarding, liking, etc. At this time, the interactive object category can be directly determined as a positive interaction category, that is, the content interaction object is interested in the target promotion content.

[0118] In step S320 of some embodiments, the extracted content interaction text may be further segmented, which is the process of splitting the text into smaller content interaction words (such as words or phrases). Text segmentation can help the system better understand the constituent elements in the text, making subsequent analysis more effective.

[0119] In step S330 of some embodiments, further, the content interaction words after word segmentation are vectorized. Word vectorization is a method of converting words into digital vectors, which enables computers to understand and process text data. Generally, common models include Word2Vec, GloVe, etc., and word vectors can capture the semantic relationship between words.

[0120] In steps S340 to S360 of some embodiments, the present application may obtain preset word vectors for preset category keywords from a predefined vocabulary. Preset category keywords refer to preset keywords that represent positive interaction categories and negative categories, and the preset category keywords and corresponding interaction categories are stored in the vocabulary. Furthermore, key features of the interactive word vectors may be extracted, that is, principal component analysis or other feature selection methods may be used to find the most meaningful part for classification or analysis from higher-dimensional word vectors to obtain interactive keyword vectors. Then, by calculating the similarity between the interactive keyword vector and the preset word vector of the preset category keyword, the degree of similarity between the two in the semantic space is determined. This is usually quantified using methods such as cosine similarity and Euclidean distance.

[0121] In step S370 of some embodiments, the interaction category of the preset category keyword with the highest word vector similarity is further used as the interaction object category of the content interaction sub-data.

[0122] In the above embodiments, the present application analyzes the content interaction text of the content interaction sub-data, extracts keywords and performs vectorization processing, and further analyzes the emotions and preferences of the content interaction objects towards the target promotion content, thereby providing effective data support for content interaction optimization and improving promotion effects.

[0123] In step S140 of some embodiments, in order to determine whether the content interaction object is an object suitable for recommendation to the target object, the content interaction object needs to be further evaluated. Specifically, the candidate object features of the content interaction object and the historical interaction data of the content interaction object in the application can be obtained first. Among them, the candidate object features can represent the basic feature information of the content interaction object, and these features may include geographic location information, interests and hobbies, etc. Historical interaction data refers to the historical interaction data of the content interaction object with other content in the application in the past.

[0124] In step S150 of some embodiments, further, the preference matching score between the content interaction object and the target promotion content can be determined based on the historical interaction data, that is, the matching degree is quantified by calculating a score, and the higher the score, the better the matching degree. The preference matching score can be a probability value (i.e., any value in the range of 0-1) or a score to reflect the preference matching degree between the content interaction object and the target promotion content.

[0125] See also Figure 4 , Figure 4 It is an optional flow chart of step S150 provided in an embodiment of the present application. In some embodiments, step S150 may specifically include steps S410 to S440:

[0126] Step S410, obtaining the associated object data of the content interaction object from the historical interaction data;

[0127] Step S420, determining the total number of associated objects, the total number of objects with the same attribute, and the total number of commonly associated objects of the content interactive object based on the associated object data;

[0128] Step S430, determining the interaction frequency and main interaction type between the content interaction object and the target object in the application based on the content interaction sub-data;

[0129] Step S440, based on the total number of associated objects, the total number of objects with the same attribute, the number of commonly associated objects, the interaction frequency and the main interaction type, determine the preference matching score between the content interaction object and the target promotion content.

[0130] In step S410 of some embodiments, the associated object data refers to object data associated with the content interactive object in the application, and may include object data followed by the content interactive object, or object data following the content interactive object. Through these data, the network relationship of the content interactive object can be constructed.

[0131] In step S420 of some embodiments, the total number of associated objects is used to represent the number of objects associated with the content interactive object in the application, the total number of objects with the same attributes is used to represent the number of objects with the same preset promotion category associated with the content interactive object in the application, and the total number of commonly associated objects is used to represent the number of objects commonly associated with the content interactive object and the target object in the application. For example, in the application, the account of the content interactive object follows a total of 20 accounts, of which there are 10 accounts in the entrepreneurial category (and 3 accounts that are commonly followed by the target object), 5 accounts in the food category, and 5 other accounts without fixed categories. At this time, the total number of associated objects of the content interactive object is 10. Since the preset promotion category is entrepreneurial, the total number of objects with the same attributes is 5, and the total number of commonly associated objects is 3.

[0132] In step S430 of some embodiments, further, the content interaction sub-data (such as historical interaction records, comments, likes, etc.) can be analyzed to obtain the number of interactions between the content interaction object and the target object within a period of time, and the main interaction form between the content interaction object and the target object (such as the most frequently used interaction types such as comments, sharing, likes, views, etc.), that is, the interaction frequency and main interaction type between the content interaction object and the target object in the application are obtained.

[0133] In step S440 of some embodiments, after determining the five dimensions of the total number of associated objects, the total number of objects with the same attributes, the number of commonly associated objects, the interaction frequency, and the main interaction type, because the present application will pre-set the corresponding intervals and scores for the indicators of each dimension, the scores corresponding to the indicators of each dimension can be determined, thereby determining the preference matching scores between the content interaction objects and the target promotion content.

[0134] It should be noted that the present application is not limited to these five dimensions and can be flexibly adjusted according to actual needs to obtain a more accurate preference matching score.

[0135] In the above embodiments, the present application can jointly determine the preference matching score based on the total number of associated objects, the total number of objects with the same attributes, the number of commonly associated objects, the interaction frequency and the main interaction type, which can better understand the behavioral characteristics and preferences of the content interaction objects, improve the accuracy of determining the preference matching score, and thus improve the accuracy of object recommendation.

[0136] See also Figure 5 , Figure 5 It is an optional flow chart of step S440 provided in an embodiment of the present application. In some embodiments, step S440 may specifically include steps S510 to S560:

[0137] Step S510, determining a first matching score of the content interaction object based on the total number of associated objects;

[0138] Step S520, determining a second matching score of the content interaction object based on the total number of objects with the same attribute;

[0139] Step S530, determining a third matching score of the content interaction object based on the number of commonly associated objects;

[0140] Step S540, determining a fourth matching score of the content interaction object based on the interaction frequency;

[0141] Step S550, determining a fifth matching score of the content interaction object based on the main interaction type;

[0142] Step S560, determining a preference matching score between the content interaction object and the target promotion content based on the first matching score, the second matching score, the third matching score, the fourth matching score and the fifth matching score of the content interaction object.

[0143] In some embodiments of the present invention, in steps S510 to S560, the indicator of the total number of associated objects can be divided into three intervals, wherein the score corresponding to [0,30) is 5, the score corresponding to [30,80) is 10, and the score corresponding to [80,+∞) is 20. Thus, by determining to which interval the total number of associated objects of the content interaction object belongs, the corresponding first matching score can be determined. The indicator of the total number of objects with the same attributes can be divided into three intervals, wherein the score corresponding to [0,10) is 5, the score corresponding to [10,20) is 10, and the score corresponding to [20,+∞) is 20. Thus, by determining to which interval the total number of objects with the same attributes of the content interaction object belongs, the corresponding second matching score can be determined. Similarly, the third matching score, the fourth matching score, and the fifth matching score can be determined, and these scores can be added up to determine the preference matching score between the content interaction object and the target promotion content.

[0144] In step S160 of some embodiments, after determining the interactive object category, candidate object features, and preference matching score, a prediction model can be used to estimate the possibility of the content interactive object being converted into a preset promotion category in the future, i.e., the target conversion score. The target conversion score is used to characterize the degree to which the content interactive object is converted into an object that matches the preset promotion category. For example, in the intelligent sales recruitment scenario of financial technology, if the preset promotion category is insurance sales entrepreneurship, the target conversion score can characterize the possibility of the content interactive object being converted into insurance sales.

[0145] See also Figure 6 , Figure 6 It is an optional flow chart of step S160 provided in an embodiment of the present application. In some embodiments, step S160 may specifically include steps S610 to S630:

[0146] Step S610, determining interaction strategy data for the content interaction object based on the interaction object category and the content interaction sub-data;

[0147] Step S620, obtaining object feedback data of the content interaction object to the interaction strategy data;

[0148] Step S630, determining the target conversion score of the content interactive object based on the interactive object category, candidate object characteristics, preference matching score and object feedback data.

[0149] In step S610 of some embodiments, the interactive feedback data refers to the feedback action based on the interaction of the content interactive object with the target promotion content. In order to improve the accuracy of the interactive strategy data, the present application will determine the interactive strategy data for the content interactive object based on the interactive object category and the content interactive sub-data. For example, if the interactive object category is a positive interactive category, then the content interactive sub-data can be analyzed, and if the content interactive sub-data indicates a desire for further communication to understand the details, then the content interactive object can be automatically followed or specific contact information can be added. If the interactive object category is a negative interactive category, then the interactive strategy data may not be generated.

[0150] In step S620 and step S630 of some embodiments, further, object feedback data of the content interaction object to the interaction strategy data is obtained. The object feedback data represents the feedback made by the content interaction object after the operation corresponding to the received interaction strategy data, for example, the communication record with the target object in the subsequent communication process. Further, based on the interaction object category, candidate object characteristics, preference matching score and object feedback data, the target conversion score of the content interaction object is determined.

[0151] It should be noted that determining the target conversion score of the content interactive object based on the interactive object category, candidate object characteristics, preference matching score and object feedback data can specifically include: obtaining corresponding scores by respectively performing scoring tests based on the interactive object category, candidate object characteristics and object feedback data, and then adding the obtained multiple scores with the preference matching score to obtain the target conversion score.

[0152] In step S170 of some embodiments, after obtaining the target conversion score, the target object may be recommended by ranking the target conversion score.

[0153] See also Figure 7 , Figure 7 It is an optional flow chart of step S170 provided in an embodiment of the present application. In some embodiments, step S170 may specifically include steps S710 to S730:

[0154] Step S710, performing level matching based on the target conversion score and the preset score level to obtain the recommended object level of the content interaction object;

[0155] Step S720, sorting the content interaction objects based on the recommendation object levels and target conversion scores to obtain a recommendation object sequence;

[0156] Step S730 , selecting a target recommended object from the recommended object sequence based on a preset recommendation quantity, and recommending the target recommended object to the target object.

[0157] In step S710 of some embodiments, considering that the number of content interaction objects may be large, the present application may perform level matching on the content interaction objects of the positive interaction category based on the scoring card method, so that the content interaction objects with scores in the scoring interval corresponding to the same scoring level are considered to be in the same group. For example, the preset scoring levels are divided into four categories, namely S, A, B and C, where the scoring interval corresponding to the S scoring level is [0,60), the scoring interval corresponding to the A scoring level is [60,80), the scoring interval corresponding to the B scoring level is [80,90), and the scoring interval corresponding to the C scoring level is [90,100].

[0158] It should be noted that the number of preset scoring levels and the corresponding scoring ranges can be flexibly adjusted according to actual scoring rules and needs without specific limitations.

[0159] In step S720 of some embodiments, after the recommendation object level is determined, each recommendation object level may also include target conversion scores of multiple content interaction objects, which can be further sorted based on the target conversion scores of the content interaction objects to obtain a recommendation object sequence.

[0160] In step S730 of some embodiments, further, when the target object needs to contact the content interactive object to complete a task corresponding to a preset promotion category, the target recommended object can be selected from the sequence of recommended objects corresponding to the highest preset rating level according to the rating from high to low, based on the preset number of recommendations (for example, the first five, the first ten, etc.) determined based on the task corresponding to the preset promotion category, and recommended to the target object.

[0161] In the above embodiments, the embodiment of the present application ensures that the content interactive object can obtain the most suitable recommended object through the entire process from target conversion scoring to recommended objects and finally to final recommendation, thereby improving the accuracy and efficiency of object recommendation, thereby improving conversion rate and user satisfaction.

[0162] It should be noted that the non-Company's software tools or components that appear in the embodiments of the present application are merely examples and do not represent actual use.

[0163] An object recommendation method provided in an embodiment of the present application, by designing an object recommendation method based on the combination of RPA and NLP, can, on the one hand, automatically interact with the associated objects of the target object through RPA to improve the interaction efficiency with key objects; on the other hand, quickly generate the target promotion content required by the target object through a content generation model built based on ChatGLM, and predict some data based on NLP to obtain the target conversion score of the content interaction object, so that the target object can give priority to information exchange with the content interaction object with a higher target conversion score, thereby improving the object conversion efficiency. Therefore, compared with the related art that does not consider various factors related to the preset promotion category when selecting objects, the present application can fully consider various factors related to the target promotion content and the content interaction object itself by determining the preference matching score and the target conversion score, thereby improving the accuracy and efficiency of object recommendation.

[0164] See also Figure 8 The present application also provides an object recommendation device, which can implement the above object recommendation method, and the device includes:

[0165] A first acquisition module 810 is used to acquire content interaction data obtained by a target object interacting with a target promotion content in an application, where the target promotion content includes a preset promotion category;

[0166] An object acquisition module 820, configured to acquire a content interaction object of a target object based on the content interaction data, wherein the content interaction data includes content interaction sub-data of the content interaction object;

[0167] A category detection module 830 is used to perform object category detection on the content interaction object based on the content interaction sub-data to obtain an interaction object category, where the interaction object category is used to indicate the emotional direction of the interaction between the content interaction object and the target promotion content in the application;

[0168] A second acquisition module 840 is used to acquire candidate object features of the content interaction object and historical interaction data of the content interaction object in the application;

[0169] A score determination module 850 is used to determine a preference matching score between a content interaction object and a target promotion content based on historical interaction data;

[0170] The object conversion module 860 is used to predict the object conversion based on the interactive object category, the candidate object characteristics and the preference matching score, and obtain the target conversion score of the content interactive object. The target conversion score is used to represent the degree to which the content interactive object is converted into an object matching the preset promotion category;

[0171] The recommendation module 870 is used to recommend objects to target objects based on the target conversion score.

[0172] The specific implementation of the object recommendation device is basically the same as the specific implementation of the above-mentioned object recommendation method, and will not be repeated here.

[0173] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned object recommendation method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0174] See also Fig. 9 , Fig. 9 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0175] The processor 910 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0176] The memory 920 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 920 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 920, and the processor 910 calls and executes the object recommendation method of the embodiment of this application;

[0177] Input / output interface 930, used to implement information input and output;

[0178] Communication interface 940, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0179] bus 950 , which transmits information between the various components of the device (e.g., processor 910 , memory 920 , input / output interface 930 , and communication interface 940 );

[0180] The processor 910 , the memory 920 , the input / output interface 930 , and the communication interface 940 are connected to each other in communication within the device via a bus 950 .

[0181] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the object recommendation method is implemented.

[0182] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0183] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0184] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0185] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0186] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0187] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0188] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0189] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0190] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0191] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0192] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0193] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. An object recommendation method, characterized in that: The method comprises: Acquire content interaction data obtained by a target object interacting with target promotion content in an application, where the target promotion content includes a preset promotion category; Acquire a content interaction object of the target object based on the content interaction data, wherein the content interaction data includes content interaction sub-data of the content interaction object; Performing object category detection on the content interaction object based on the content interaction sub-data to obtain an interaction object category, wherein the interaction object category is used to indicate an emotional direction of the interaction between the content interaction object and the target promotion content in the application; Acquire candidate object features of the content interactive object and historical interaction data of the content interactive object in the application; Determine a preference matching score between the content interaction object and the target promotion content based on the historical interaction data; Performing object conversion prediction based on the interactive object category, the candidate object features and the preference matching score to obtain a target conversion score for the content interactive object, wherein the target conversion score is used to characterize the degree to which the content interactive object is converted into an object matching the preset promotion category; An object is recommended to the target object based on the target conversion score.

2. The method according to claim 1, characterized in that The recommending an object to the target object based on the target conversion score includes: Performing grade matching based on the target conversion score and the preset score level to obtain the recommended object level of the content interaction object; Sort the content interaction objects based on the recommendation object levels and the target conversion scores to obtain a recommendation object sequence; A target recommended object is selected from the recommended object sequence based on a preset recommendation quantity, and the target recommended object is recommended to the target object.

3. The method according to claim 1, characterized in that The step of determining the preference matching score between the content interaction object and the target promotion content based on the historical interaction data includes: Acquire the associated object data of the content interaction object from the historical interaction data; Determine the total number of associated objects, the total number of objects with the same attributes, and the total number of commonly associated objects of the content interactive object based on the associated object data, the total number of associated objects being used to represent the number of objects associated with the content interactive object in the application, the total number of objects with the same attributes being used to represent the number of objects with the same preset promotion category associated with the content interactive object in the application, and the total number of commonly associated objects being used to represent the number of objects commonly associated between the content interactive object and the target object in the application; Determine the interaction frequency and main interaction type between the content interaction object and the target object in the application based on the content interaction sub-data; Based on the total number of associated objects, the total number of objects with the same attribute, the number of commonly associated objects, the interaction frequency and the main interaction type, the preference matching score between the content interaction object and the target promotion content is determined.

4. The method according to claim 3, characterized in that The determining the preference matching score between the content interaction object and the target promotion content based on the total number of associated objects, the total number of objects with the same attribute, the number of commonly associated objects, the interaction frequency, and the main interaction type includes: Determining a first matching score of the content interaction object based on the total number of associated objects; Determining a second matching score of the content interaction object based on the total number of objects with the same attribute; Determining a third matching score of the content interaction object based on the number of commonly associated objects; Based on the interaction frequency, determining a fourth matching score of the content interaction object; Based on the primary interaction type, determining a fifth matching score of the content interaction object; The preference matching score between the content interactive object and the target promotion content is determined based on the first matching score, the second matching score, the third matching score, the fourth matching score and the fifth matching score of the content interactive object.

5. The method according to claim 1, characterized in that Before obtaining the content interaction data obtained by the target object interacting with the target promotion content in the application, the method further includes: Get the object type of the target object and generate content conditional text; Performing content generation on the object type and the generated content condition text based on a preset content generation model to obtain candidate promotion content, wherein the candidate promotion content is stored in a content library of the content generation model; Select target promotion content from the candidate promotion content, and mark the target promotion content in the content library to update the priority of the target promotion content in the content library.

6. The method according to any one of claims 1 to 5, characterized in that: The performing object category detection on the content interactive object based on the content interactive sub-data to obtain the interactive object category includes: Acquire content interaction text from the content interaction sub-data; Performing text segmentation on the content interaction text to obtain content interaction words; Performing word vectorization on the content interaction words to obtain interaction word vectors; Get the preset word vector of the preset category keyword; Extract key features of the interactive word vector to obtain an interactive keyword vector; Performing vector similarity calculation on the preset word vector and the interactive keyword vector to obtain word vector similarity; The interactive object category is determined based on the word vector similarity.

7. The method according to any one of claims 1 to 5, characterized in that: The object conversion prediction is performed based on the interactive object category, the candidate object features and the preference matching score to obtain the target conversion score of the content interactive object, including: Determine interaction strategy data for the content interaction object based on the interaction object category and the content interaction sub-data; Obtain object feedback data of the content interaction object to the interaction strategy data; The target conversion score of the content interactive object is determined based on the interactive object category, the candidate object characteristics, the preference matching score and the object feedback data.

8. An object recommendation device, characterized in that: The device comprises: A first acquisition module is used to acquire content interaction data obtained by a target object interacting with a target promotion content in an application, wherein the target promotion content includes a preset promotion category; An object acquisition module, configured to acquire a content interaction object of the target object based on the content interaction data, wherein the content interaction data includes content interaction sub-data of the content interaction object; A category detection module, configured to perform object category detection on the content interaction object based on the content interaction sub-data to obtain an interaction object category, wherein the interaction object category is used to indicate an emotional direction of the interaction between the content interaction object and the target promotion content in the application; A second acquisition module, used to acquire candidate object features of the content interactive object and historical interaction data of the content interactive object in the application; A score determination module, used to determine a preference matching score between the content interaction object and the target promotion content based on the historical interaction data; An object conversion module, configured to predict object conversion based on the interactive object category, the candidate object features and the preference matching score, and obtain a target conversion score for the content interactive object, wherein the target conversion score is used to characterize the degree to which the content interactive object is converted into an object matching the preset promotion category; A recommendation module is used to recommend an object to the target object based on the target conversion score.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images