Personalized advertisement putting method and device, equipment and storage medium

Abstract

The invention discloses a personalized advertisement putting method and device, equipment and a storage medium, and relates to the technical field of intelligent advertisements, and the personalized advertisement putting method comprises the steps: obtaining user interaction data, initial advertisement content and sensitive word classification; processing the user interaction data according to the sensitive word classification through a multi-modal agent to obtain a target sensitive word; performing optimization adjustment on the initial advertisement content according to the multi-modal intelligent agent and the target sensitive word to obtain target advertisement content; and pushing and displaying the target advertisement content to complete advertisement putting. According to the method and the device, accurate advertisement putting is realized, the effectiveness of the advertisement and the experience satisfaction degree of the user are also improved, the advertisement content received by the user better conforms to personal interests and requirements of the user, the annoyance of the advertisement talking over the market is avoided, and the user experience is improved. Therefore, the maximization of the advertisement effect and the optimization of the user experience are realized in the whole advertisement putting process.

Description

Technical Field

[0001] The present application relates to the field of intelligent advertising technology, and in particular to a personalized advertising delivery method, device, equipment and storage medium. Background Art

[0002] With the rapid development of information technology, Internet advertising has become an important means for companies to promote their products and services. In the era of digital marketing, personalized and accurate advertising has become particularly important. Users' demands for advertising content are becoming increasingly diverse, and they expect to receive advertising information that matches their interests and needs. Therefore, advertising delivery methods need to be able to fully consider the personalized needs of users, especially those user groups who are sensitive to advertising, in order to increase the attractiveness of advertising and user satisfaction.

[0003] At present, traditional advertising methods are mainly based on an inherent crowd label system, which is derived from the statistical data of users' browsing, clicking, purchasing and other behaviors of products. Although this method can achieve targeted advertising to a certain extent, it usually ignores the personalized needs of users, especially their sensitivity to advertising content. Advertising content is often too general and difficult to meet the specific needs of users, which leads to users' boredom with advertising and reduces the effectiveness of advertising.

[0004] The main problem with existing advertising delivery methods is the lack of in-depth understanding and real-time response to users' personalized needs. Since users' sensitive words and personalized preferences are not fully considered, advertising content often cannot accurately match users' actual needs, resulting in poor advertising delivery results. In addition, traditional methods cannot capture changes in users' sensitive words in real time, resulting in delayed updates of advertising content and failure to reflect users' latest preferences in a timely manner. These problems ultimately lead to reduced advertising effectiveness and reduced user experience satisfaction. Therefore, how to improve advertising effectiveness and user experience satisfaction has become an urgent problem to be solved.

[0005] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The purpose of this application is to provide a personalized advertising delivery method, device, equipment and storage medium, aiming to solve the technical problem of how to improve the effectiveness of advertising and user experience satisfaction.

[0007] To achieve the above objectives, the present application proposes a personalized advertising delivery method, the method comprising:

[0008] Obtain user interaction data, initial advertising content, and sensitive word classification;

[0009] Processing the user interaction data according to the sensitive word classification by a multimodal agent to obtain target sensitive words;

[0010] Optimizing and adjusting the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content;

[0011] The target advertisement content is pushed and displayed to complete the advertisement delivery.

[0012] In one embodiment, the step of processing the user interaction data according to the sensitive word classification by a multimodal agent to obtain a target sensitive word includes:

[0013] Extracting initial sensitive words from the user interaction data according to the sensitive word classification by a multimodal agent;

[0014] The sensitive word score is calculated according to the preset first-level classification frequency, the preset first-level classification weight, the preset second-level classification frequency, the preset second-level classification weight, the initial sensitive word and the preset sensitive word weight;

[0015] The initial sensitive words are filtered according to the sensitive word scores to obtain target sensitive words.

[0016] In one embodiment, the multimodal agent is constructed based on a multimodal macro model and advertising prompt words, and the step of optimizing and adjusting the initial advertising content based on the multimodal agent and the target sensitive words to obtain the target advertising content includes:

[0017] Constructing an optimized prompt word according to the target sensitive word, the advertising prompt word and the initial advertising content;

[0018] The multimodal intelligent agent performs semantic analysis and content optimization on the initial advertisement content according to the optimization prompt words and the target sensitive words to obtain target advertisement content.

[0019] In one embodiment, the step of performing semantic analysis and content optimization on the initial advertisement content according to the optimization prompt words and the target sensitive words by the multimodal agent to obtain the target advertisement content includes:

[0020] Performing semantic analysis on the initial advertisement content by the multimodal agent to obtain original semantics;

[0021] The multimodal agent integrates the target sensitive words into the initial advertisement content according to the optimized prompt words and the original semantics to obtain the target advertisement content.

[0022] In one embodiment, after the step of pushing and displaying the personalized advertisement and completing the step of advertisement delivery, the process further includes:

[0023] Get user feedback;

[0024] The preset sensitive word weight is adjusted according to the user feedback.

[0025] In one embodiment, the step of pushing and displaying the target advertisement content to complete advertisement delivery includes:

[0026] Determining audience characteristics and audience preferences based on the user interaction data and the target sensitive words;

[0027] Determine a target advertising platform and advertising push parameters according to the audience characteristics and the audience preferences;

[0028] Obtaining the format requirements and review requirements of the target advertising platform;

[0029] Adjust the target advertisement content according to the format requirement and the review requirement to obtain the advertisement content to be pushed;

[0030] The advertisement content to be pushed is pushed and displayed according to the advertisement push parameters to complete the advertisement delivery.

[0031] In one embodiment, the step of obtaining user interaction data, initial advertisement content, and sensitive word classification includes:

[0032] Collect user interaction data through the preset platform;

[0033] Obtaining user features according to the user interaction data;

[0034] Acquire initial advertisement content from a preset advertisement database according to the user characteristics;

[0035] Sensitive words are classified based on market research, user surveys and product feature definitions.

[0036] In addition, to achieve the above-mentioned purpose, the present application also proposes a personalized advertising delivery device, the device comprising:

[0037] Data acquisition module, used to obtain user interaction data, initial advertising content and sensitive word classification;

[0038] A data processing module, configured to process the user interaction data according to the sensitive word classification through a multimodal agent to obtain target sensitive words;

[0039] An advertisement optimization module, configured to optimize and adjust the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content;

[0040] The advertisement push module is used to push and display the target advertisement content to complete the advertisement delivery.

[0041] In addition, to achieve the above-mentioned purpose, the present application also proposes a personalized advertising delivery device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the personalized advertising delivery method described above.

[0042] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the personalized advertising delivery method described above are implemented.

[0043] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the personalized advertising delivery method described above are implemented.

[0044] One or more technical solutions proposed in this application have at least the following technical effects:

[0045] Obtain user interaction data, initial advertising content, and sensitive word classification; process the user interaction data according to the sensitive word classification through a multimodal agent to obtain target sensitive words; optimize and adjust the initial advertising content according to the multimodal agent and the target sensitive words to obtain target advertising content; push and display the target advertising content to complete advertising delivery. First, the system obtains the user's behavior patterns and preferences by collecting and analyzing user interaction data, and determines the initial advertising content and sensitive word classification, which provides necessary data support for advertising personalization. Then, the multimodal agent is used to process these data, and the target sensitive words are extracted according to the sensitive word classification, accurately capturing the user's specific interests and potential needs, laying the foundation for personalized adjustment of advertising content. Then, the system optimizes and adjusts the initial advertising content according to the analysis results of the multimodal agent and the target sensitive words, and generates target advertising content that better meets the user's personalized needs, thereby improving the relevance and attractiveness of the advertising content, making it more likely to arouse the interest and response of the target user group. Finally, these optimized target advertising contents are pushed to the target users through appropriate channels and displayed on the user's device or platform to complete the entire advertising delivery process. This not only achieves precise advertising delivery and improves advertising conversion rate and return on investment, but also improves advertising effectiveness and user experience satisfaction, because the advertising content received by users is more in line with their personal interests and needs, avoiding boredom with general advertisements, thereby maximizing advertising effectiveness and optimizing user experience throughout the entire advertising delivery process. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 A flowchart of the first embodiment of the personalized advertising delivery method of the present application is provided;

[0049] Figure 2 A flowchart of the second embodiment of the personalized advertising delivery method of the present application is provided;

[0050] Figure 3 A brief flowchart of the personalized advertising delivery method provided in Example 2 of the present application;

[0051] Figure 4 This is a schematic diagram of the module structure of the personalized advertising delivery device according to an embodiment of the present application;

[0052] Figure 5 Schematic diagram of the device structure of the hardware operating environment involved in the personalized advertising delivery method in the embodiment of the present application.

[0053] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0055] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0056] With the advancement of information technology, Internet advertising has become an important means of corporate promotion. However, in the era of digital marketing, the traditional advertising method based on fixed tags ignores the personalized needs of users and the sensitivity of advertising, resulting in generalized advertising content and failure to accurately match the actual needs of users, thereby reducing advertising appeal and user experience. These problems stem from the fact that existing methods lack a deep understanding of user needs and the ability to respond in real time, which ultimately affects advertising effectiveness and user satisfaction.

[0057] The main solution of the embodiment of the present application is: the system determines behavior patterns and preferences by collecting and analyzing user data, uses multimodal intelligent agents to extract sensitive words and adjust advertising content to make it more in line with user needs, and then pushes the optimized advertisements through appropriate channels, thereby improving advertising conversion rates and user experience satisfaction.

[0058] It should be noted that the execution subject of the embodiment of the present application can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, an advertising push system, etc. The following takes the advertising push system as an example to illustrate this embodiment and the following embodiments.

[0059] Based on this, the embodiment of the present application provides a personalized advertising delivery method, referring to Figure 1 , Figure 1 This is a flowchart of the first embodiment of the personalized advertising delivery method of the present application.

[0060] In this embodiment, the personalized advertising delivery method includes steps S10 to S40:

[0061] Step S10, obtaining user interaction data, initial advertisement content, and sensitive word classification;

[0062] It should be noted that user interaction data refers to data generated by various interactive behaviors of users related to advertisements or products on the Internet, including but not limited to user clicks, comments, favorites, sharing and purchases, which reflect the user's interest and preference for specific advertisements or products. These interactive data are used to extract user sensitive words through multimodal technology to help optimize advertising content to make it more in line with the user's actual needs and preferences. The types of user interaction data can include text, image and video information. Initial advertising content refers to the advertising text, pictures, videos or other forms of advertising materials pre-set in the advertising system before personalized optimization. These contents are usually formulated based on brand marketing strategies and product features to attract the attention of potential customers. The initial advertising content will be combined with sensitive words extracted from user interaction data and optimized through large model technology to generate advertising display content that is more in line with user preferences. Sensitive word classification refers to the classification of sensitive words extracted from user interaction data according to different dimensions and categories. Sensitive words are divided into multiple categories, such as semantic categories, situational categories, user group categories, psychology and behavior categories, commodity category-related sensitive words, cultural and regional categories, technology and innovation categories, lifestyle categories, sensory experience categories, psychological state and motivation categories, economic status and consumption capacity categories, environmental protection and social responsibility categories, and time urgency categories. These classifications help the advertising system understand user preferences more carefully and match user needs more accurately during the optimization of advertising content. In this way, advertising content can be personalized to improve the relevance and attractiveness of advertising, thereby improving user experience and advertising effectiveness.

[0063] It is understandable that user interaction data is first obtained by tracking the user's interaction records with product copy and purchased goods. This step involves collecting user behavior data on the network platform, such as clicks, comments, favorites, sharing and purchases, etc. These data types include text, image and video information. The purpose of this is to capture the user's interests and preferences, because these behavioral data can intuitively reflect the user's attention and response to specific content. Next, the initial advertising content refers to the advertising materials pre-set in the advertising system before personalized optimization. These materials are usually designed based on brand strategy and product characteristics to attract the attention of potential customers. Then, sensitive word classification refers to the classification of sensitive words extracted from user interaction data according to different dimensions, such as functionality, emotion, brand reputation, etc. This step extracts sensitive words from the user's daily interaction content by constructing a specific prompt structure, using multimodal large models and natural language processing technology, and stores them in the database by category. The effect of this is to enable the advertising system to understand the user's preferences more carefully and match the user's needs more accurately in the process of advertising content optimization, ultimately achieving personalized adaptation of advertising content and improving user experience and advertising effectiveness. In this way, advertising can prevent users from getting bored with general advertising content, thereby increasing the click-through rate and conversion rate of ads, while also enhancing user satisfaction and brand loyalty.

[0064] As an example, the steps of obtaining user interaction data, initial advertising content and sensitive word classification include: collecting user interaction data through a preset platform; obtaining user characteristics based on the user interaction data; obtaining initial advertising content from a preset advertising database based on the user characteristics; and obtaining sensitive word classification based on market research, user surveys and product feature definitions.

[0065] A preset platform refers to a network platform or service environment that is specially set up to collect user interaction data. These platforms may include social media, e-commerce websites, mobile applications, search engines, etc. For example, an e-commerce website may preset tracking codes to collect data on user browsing, searching, adding to shopping carts, and purchasing goods. User characteristics are a series of labels or attributes that describe user preferences, behavioral habits, and needs based on the analysis of user interaction data. These characteristics may include user interest in a certain type of product, purchasing power, brand preference, active time period, etc. For example, a user characteristic may be "a middle-aged woman who has a high interest in high-end skin care products and tends to browse related products at night." A preset advertising database refers to a database that stores various advertising materials and content. These advertising materials can be in the format of text, pictures, videos, or audio. They are prepared in advance for different user groups or different marketing activities. For example, a preset advertising database may contain advertising templates designed for different festivals, different user groups, or different product features. Market research, user research, and product feature definition refer to the process of determining advertising strategies and sensitive word classification through market analysis, user surveys, and product feature analysis. Market research may involve analyzing industry trends, competitor advertising strategies, etc.; user research may include questionnaires, focus group discussions, etc. to understand user needs and preferences; product feature definition is to clarify the unique selling points and target markets of products. This information is combined to help define sensitive word classification to ensure that advertising content matches users' actual needs and market trends. For example, market research may find that "environmental protection" is a hot trend at present, user research may show that the target user group has a high degree of concern for "natural ingredients", and product feature definition may emphasize "sustainable production" and "no chemical additives". This information together guides the setting of sensitive word classification.

[0066] First, through preset platforms such as social media, e-commerce websites or mobile applications, the interactive data generated by users' browsing, searching, commenting, collecting, sharing and purchasing behaviors are collected; these data are used to analyze and summarize user characteristics, that is, users' interests, preferences and behavior patterns. For example, a user who often browses outdoor equipment and purchases related products may be marked as an "outdoor activity enthusiast". Then, based on these user characteristics, the system will filter out the initial advertising content that matches them from the preset advertising database, which contains a variety of advertising materials prepared in advance for different user groups or different marketing activities. Finally, the sensitive word classification is based on market research, user surveys and product feature definitions. These studies may include analyzing industry trends, user needs, competitor strategies and unique selling points of products to ensure that advertising content can accurately touch users' needs and interests. For example, if market research shows that environmental protection is the focus of users' concern, and the product feature definition emphasizes the use of sustainable materials, then "environmental protection" may become a sensitive word classification used to optimize advertising content to make it more attractive to the target user group. Through this series of steps, the advertising delivery system can provide more personalized and targeted advertising content to improve the effectiveness of advertising and user experience satisfaction.

[0067] Step S20, processing the user interaction data according to the sensitive word classification by a multimodal agent to obtain target sensitive words;

[0068] It should be noted that a multimodal agent refers to an artificial intelligence system that can process and analyze multiple types of data (such as text, images, videos, etc.). This agent is usually based on advanced machine learning models and can understand and extract information from different modal data and use it for decision-making or prediction. In the context of advertising, a multimodal agent uses the constructed prompt structure to extract sensitive words from the user's daily interactive content. These sensitive words reflect the user's interests, preferences and possible sensitive points. The agent uses technologies such as natural language processing, image recognition and video analysis to identify and classify these sensitive words for subsequent advertising content optimization. Target sensitive words refer to keywords or phrases that are extracted from user interaction data and are closely related to user characteristics. These sensitive words are selected according to the sensitive word classification system and can represent the user's needs, preferences or sensitivity to certain advertising content. For example, if a user often buys organic food and shows a high degree of concern for environmental protection topics, then "organic" and "environmental protection" may become the user's target sensitive words. These target sensitive words are then used to generate personalized advertising content to ensure that the advertisements can more accurately attract the target user groups and improve the relevance and attractiveness of the advertisements.

[0069] It can be understood that, first, the multimodal agent analyzes the interactive data such as text, pictures and videos generated by users on different platforms through integrated algorithms and models, such as natural language processing (NLP) and computer vision technology; secondly, the agent identifies and extracts keywords or phrases related to user behavior and preferences based on the predefined sensitive word classification system. These classifications may include multiple dimensions such as functionality, emotion, brand reputation, etc.; finally, the "target sensitive words" obtained through this process, that is, those words that best represent the user's personalized needs and preferences, will be used to generate and optimize personalized advertising content. The effect of this is to make the advertising content more in line with the user's actual needs and interests, improve the personalization level and relevance of the advertisement, thereby improving the effectiveness of the advertisement and increasing user engagement and satisfaction.

[0070] Step S30, optimizing and adjusting the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content;

[0071] It should be noted that targeted advertising content refers to the final advertising materials that have been optimized and adjusted, and these materials have been personalized according to the sensitive words and preferences of the target users. This content is designed to improve the relevance of the advertisement and make it more attractive to specific user groups, thereby improving the effectiveness of the advertisement and the user's experience satisfaction. Specifically, targeted advertising content includes advertising slogans, images, videos and other elements that have been adjusted according to user sensitive words to ensure that the advertising information can resonate with the target users and prompt them to take actions, such as clicking, purchasing or participating in brand activities.

[0072] It is understandable that the multimodal agent first identifies and understands the semantic and emotional tendencies of the initial advertising content, and then fine-tunes the advertising content in combination with the target sensitive words extracted from the user interaction data. This fine-tuning may include replacing or adding vocabulary, changing sentence structure, or adjusting visual elements to ensure that the advertising content not only conveys the original marketing information, but also is closer to the user's personalized needs and preferences. For example, if the target sensitive words include "environmental protection" and "natural", the agent may highlight the environmental protection characteristics and natural ingredients of the product in the advertisement to attract users who are sensitive to these characteristics. In this way, the optimized advertising content can more effectively attract the target user group and improve the attractiveness and conversion rate of the advertisement.

[0073] As an example, a multimodal agent is constructed based on a multimodal large model and advertising prompt words, and the step of optimizing and adjusting the initial advertising content based on the multimodal agent and the target sensitive words to obtain the target advertising content includes: constructing optimized prompt words based on the target sensitive words, the advertising prompt words and the initial advertising content; and performing semantic analysis and content optimization on the initial advertising content based on the optimized prompt words and the target sensitive words by the multimodal agent to obtain the target advertising content.

[0074] A multimodal big model refers to a complex artificial intelligence model that can process and understand multiple types of data (such as text, images, videos, etc.). This model is usually based on deep learning technology and can extract features from data of different modalities and conduct comprehensive analysis to identify patterns and meanings. In the field of advertising, a multimodal big model can analyze text, image, and video content in user interaction data to extract sensitive words and user preferences. An advertising prompt is a series of instructions or questions used to guide a multimodal agent to perform a specific task. In the context of advertising content optimization, an advertising prompt may contain guidance on how to extract sensitive words from user interaction data, how to understand the semantics of advertising content, and how to adjust advertising content according to user preferences. Optimized prompts are specially constructed prompts based on target sensitive words and initial advertising content during the advertising content optimization process. These prompts are used to guide multimodal agents on how to combine users' sensitive word preferences and initial advertising content to generate more personalized and attractive advertising content. Semantic analysis and content optimization refer to a series of processing steps that multimodal agents perform on the initial advertising content, including understanding the deep meaning of the advertising content (semantic analysis) and adjusting the advertising content according to target sensitive words and user preferences (content optimization). This process may involve replacing or adding vocabulary, adjusting sentence structure, changing visual elements, etc., to ensure that the optimized advertising content better attracts the target user group without changing the original meaning, and improves the relevance and attractiveness of the advertisement.

[0075] First, the multimodal agent is built on a multimodal big model that can process and analyze data from different sources and formats, such as text, images, and videos. Then, it is combined with advertising prompts, which are a series of carefully designed instructions or questions, to guide the agent to perform specific advertising content optimization tasks. Next, based on the target sensitive words, advertising prompts, and initial advertising content, optimized prompts are constructed. These prompts are specially designed to guide the agent on how to adjust the advertising content according to the user's sensitive word preferences. Finally, the multimodal agent uses these optimized prompts and target sensitive words to conduct in-depth semantic analysis of the initial advertising content, understand the deep meaning of the advertisement, and optimize the content, such as adjusting the wording, strengthening certain information points, or adjusting visual elements, to generate the final target advertising content. Such content not only conveys the core information of the advertisement, but also better fits the personalized needs and preferences of users, thereby improving the attractiveness and effectiveness of the advertisement.

[0076] As an example, the step of performing semantic analysis and content optimization on the initial advertising content according to the optimization prompt words and the target sensitive words by the multimodal agent to obtain the target advertising content includes: performing semantic analysis on the initial advertising content by the multimodal agent to obtain original semantics; and integrating the target sensitive words into the initial advertising content according to the optimization prompt words and the original semantics by the multimodal agent to obtain the target advertising content.

[0077] Original semantics refers to the original meaning and intent of the initial advertising content before any personalized optimization processing, which includes the core information, brand value, product features or promotional points that the advertisement originally wanted to convey. Original semantics is the basis for advertising content optimization. It ensures that the basic purpose and information of the advertisement will not be changed or distorted when personalized adjustments are made.

[0078] During the optimization process, the multimodal agent first performs semantic analysis on the initial ad content to identify and understand its original semantics. This step involves in-depth analysis of different modal data such as ad text, images, and videos to extract key information and concepts. Then, based on the optimization prompt words and original semantics, the agent cleverly integrates the target sensitive words into the initial ad content to generate the target ad content. The purpose of this process is to make the ad content more in line with the user's personalized needs and preferences while maintaining the original semantics, thereby improving the relevance and attractiveness of the ad. In this way, the optimized ad content can better attract the target user group and improve the effectiveness of the ad without changing the original intention.

[0079] Step S40: Push and display the target advertisement content to complete advertisement delivery.

[0080] It should be noted that push and display refer to the process of delivering optimized and adjusted target advertising content to the target user group through appropriate channels and methods, and displaying these advertising contents on the user's device or platform.

[0081] It is understandable that first, the system needs to identify the target user group that is most likely to be interested in the advertising content based on information such as user characteristics, behavioral data, and target sensitive words. Then, the system will select or generate advertising content that best matches the target user group from the advertising library, which has been personalized and optimized based on the user's sensitive words and preferences. Next, the system will formulate a suitable push strategy based on factors such as the target user's active time period, preferred platform and device, which may include selecting the appropriate advertising display location, push time and frequency, etc. Finally, the system will push the target advertising content to the user's device or platform and display these ads during the user's browsing or use. The display methods may include banner ads, video ads, native ads, social media posts, and other forms. After the advertising content is displayed, the system will collect user feedback data on the advertisement, such as click-through rate, conversion rate, viewing time, and other indicators to evaluate the effectiveness of the advertising. Based on this data, the system can further optimize the push strategy and advertising content to improve the effectiveness and ROI of the advertisement.

[0082] As an example, the target advertising content is pushed and displayed, and the steps of completing the advertising delivery include: determining audience characteristics and audience preferences based on the user interaction data and the target sensitive words; determining the target advertising platform and advertising push parameters based on the audience characteristics and the audience preferences; obtaining the format requirements and review requirements of the target advertising platform; adjusting the target advertising content according to the format requirements and the review requirements to obtain the advertising content to be pushed; pushing and displaying the advertising content to be pushed according to the advertising push parameters to complete the advertising delivery.

[0083] Audience characteristics refer to the common attributes of user groups derived from the analysis of user interaction data and target sensitive words, such as age, gender, geographic location, device usage habits, etc. Audience preferences refer to the preferences of this user group for advertising content, such as their preferences for certain themes, styles or product features. Understanding audience characteristics and preferences helps to accurately locate the target user group of the advertisement and design advertising content that is more likely to arouse their interest. Target advertising platforms refer to the most suitable channels for displaying advertisements based on audience characteristics and preferences, such as social media platforms, search engines, news websites or other online media. Advertising push parameters include the time and frequency of advertising display, the active time period of the target user group, advertising budget, bidding strategy, etc. These parameters determine when, where and how the advertisement is displayed to the target audience. Format requirements refer to the technical specifications of the target advertising platform for advertising content, such as size, resolution, file type, video length, etc. Audit requirements refer to the compliance requirements of the platform for advertising content, including that the advertising content cannot violate platform policies, laws and regulations, and cannot contain inappropriate content. These requirements ensure that the advertising content meets technical specifications and can be displayed safely and compliantly. Advertising content to be pushed refers to advertising content that has been adjusted according to the format requirements and review requirements of the target advertising platform and is ready to be pushed and displayed. This content has taken into account audience characteristics and preferences, optimized advertising materials, and complies with the technical specifications and compliance requirements of the target platform. It is the advertising version that will ultimately be displayed to the target user group.

[0084] First, by analyzing user interaction data and target sensitive words, the system can determine the characteristics and preferences of the target audience, such as their most active time periods, preferred content types, and product features of interest. Then, based on these audience characteristics and preferences, the system selects the most appropriate advertising platform and sets advertising push parameters, such as pushing ads to users of a certain age group in the evening on social media. Next, the system obtains the format requirements and review requirements of the target advertising platform for advertising content, such as the length limit of video ads and prohibited content types. After that, based on these requirements, the system makes necessary adjustments to the target advertising content, such as cropping the video to meet the length requirements or modifying sensitive words to meet the review standards, thereby obtaining the advertising content to be pushed. Finally, according to the previously set push parameters, the system displays the adjusted advertising content to be pushed to the target audience at the right time and on the right platform, completing the entire advertising delivery process. This process ensures that the advertising content can accurately reach the target user group and meets the specifications of each advertising platform to increase advertising exposure and user engagement.

[0085] This embodiment provides a personalized advertising delivery method, which obtains user interaction data, initial advertising content, and sensitive word classification; processes the user interaction data according to the sensitive word classification through a multimodal agent to obtain target sensitive words; optimizes and adjusts the initial advertising content according to the multimodal agent and the target sensitive words to obtain target advertising content; pushes and displays the target advertising content to complete advertising delivery. First, the system obtains the user's behavior pattern and preference by collecting and analyzing user interaction data, and determines the initial advertising content and sensitive word classification at the same time, which provides necessary data support for advertising personalization. Then, the multimodal agent is used to process these data, and the target sensitive words are extracted according to the sensitive word classification, accurately capturing the user's specific interests and potential needs, and laying the foundation for personalized adjustment of advertising content. Then, the system optimizes and adjusts the initial advertising content according to the analysis results of the multimodal agent and the target sensitive words, and generates target advertising content that better meets the personalized needs of users, thereby improving the relevance and attractiveness of the advertising content, making it more likely to arouse the interest and response of the target user group. Finally, these optimized target advertising contents are pushed to target users through appropriate channels and displayed on users’ devices or platforms, completing the entire advertising delivery process. This not only achieves accurate advertising delivery and improves advertising conversion rate and return on investment, but also improves advertising effectiveness and user experience satisfaction, because the advertising content received by users is more in line with their personal interests and needs, avoiding boredom with general advertisements, thus maximizing advertising effects and optimizing user experience in the entire advertising delivery process.

[0086] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can refer to the above introduction, and will not be repeated later. Figure 2 , Figure 2 This is a flow chart of the second embodiment of the personalized advertising delivery method of the present application. Step S20 of the personalized advertising delivery method includes steps S21 to S23:

[0087] Step S21, extracting initial sensitive words from the user interaction data according to the sensitive word classification by a multimodal agent;

[0088] It should be noted that initial sensitive words refer to keywords or phrases initially extracted from user interaction data through multimodal agents and sensitive word classification. These words are words that users frequently use or show sensitivity to in various interactive behaviors related to advertisements or products. They may be related to users' needs, preferences, emotional reactions, or attention to specific topics. Initial sensitive words are the starting point of the advertising content optimization process because they are directly extracted from users' original interaction data and may not have been deeply analyzed or personalized. These sensitive words can then be further analyzed and classified to determine their association with users' specific needs and preferences, and then used to optimize advertising content to make it more personalized and targeted. In short, initial sensitive words provide the basis for creating advertising content that better meets users' actual needs.

[0089] Understandably, the multimodal agent first accesses user interaction data, which is derived from the user's interaction with content on different platforms, including text, pictures, videos and other forms. The agent uses advanced natural language processing, image recognition and video analysis technologies to identify and extract initial sensitive words related to user behavior and preferences from these rich data according to the predefined sensitive word classification system. These sensitive words may involve users' attention to specific product features, emotional tendencies or discussions on certain topics, reflecting users' needs and interests. The extraction of initial sensitive words is the basis for the generation of personalized advertising content. It enables the advertising system to initially understand the user's concerns and provide key information for subsequent in-depth analysis and advertising content optimization. Through this process, the advertising system can capture user needs more accurately and lay the foundation for creating more attractive and relevant personalized advertising content.

[0090] Step S22, calculating a sensitive word score according to a preset first-level classification frequency, a preset first-level classification weight, a preset second-level classification frequency, a preset second-level classification weight, the initial sensitive word, and a preset sensitive word weight;

[0091] It should be noted that the preset first-level classification frequency, preset first-level classification weight, preset second-level classification frequency, and preset second-level classification weight refer to the preset parameters of different levels of classification in the sensitive word classification system. The first-level classification and the second-level classification refer to the hierarchical structure of sensitive words, where the first-level classification is a broader category, and the second-level classification is a more specific subcategory under the first-level classification. For example, the first-level classification may be "healthy", and its second-level classification may be "no side effects", "low sensitivity", etc. The preset frequency refers to the frequency of occurrence of sensitive words under these categories in user interaction data, and the preset weight is the importance score of these categories in the overall advertising optimization. For example, if the preset weight of the first-level classification "healthy" is set to 2, and the preset weight of its second-level classification "no side effects" is set to 1.5, this means that when calculating the sensitive word score, the appearance of "healthy" and "no side effects" will have different effects on the final sensitive word score. The preset sensitive word weight refers to the preset importance score of each sensitive word in the sensitive word classification system. For example, the word "no side effects" may be considered particularly important to certain user groups, so its preset weight may be set to a higher value, such as 1.8, to ensure that this sensitive word can more significantly affect the generation of advertising content when optimizing advertising content. The sensitive word score is a quantitative indicator calculated based on the above preset frequency and weight, which is used to measure the importance of specific sensitive words in advertising content optimization. This score reflects the frequency of occurrence of sensitive words in user interaction data and its preset weight, helping the advertising system determine which sensitive words should be given priority. For example, if a sensitive word appears frequently in user interaction data and has a high preset weight, then its score will be high, indicating that it should be highlighted in the advertising content to increase the relevance and attractiveness of the advertisement.

[0092] It is understandable that the system first scores the initial sensitive words based on the preset first-level classification frequency and weight and the second-level classification frequency and weight. The first-level classification frequency refers to the frequency of a broad category such as "health" in the user interaction data, and its weight is the importance score of this category in advertising. For example, the weight of the "health" category may be set to 2, indicating that it is an important consideration. The second-level classification frequency and weight are for more specific subcategories, such as "no side effects" or "low sensitivity", and their frequencies and weights will also affect the final sensitive word score. The initial sensitive words are keywords extracted directly from the user interaction data, and the preset sensitive word weights are the relative importance of these words in optimizing advertising content. The system combines these factors and calculates a score for each sensitive word through a calculation formula, such as score = (first-level classification frequency * first-level classification weight) + (second-level classification frequency * second-level classification weight) + (sensitive word frequency * sensitive word weight). This score reflects the relative importance of sensitive words in user preferences, and helps the advertising system determine which sensitive words should be prioritized in advertising content optimization to improve the relevance and attractiveness of the ads, ultimately achieving the goal of improving advertising effectiveness.

[0093] Step S23: filtering the initial sensitive words according to the sensitive word scores to obtain target sensitive words.

[0094] It should be noted that filtering refers to the process of screening the initial sensitive words based on the sensitive word scores. The system will set one or more thresholds based on the scores of each sensitive word calculated previously, and only sensitive words with scores exceeding these thresholds will be retained. The purpose of this filtering process is to identify and retain those sensitive words that are most likely to have an impact on users, that is, those words that best represent user preferences and interests.

[0095] It is understandable that the system will filter the initial sensitive words extracted from the user interaction data based on the sensitive word scores calculated previously. Specifically, the system will set a score threshold, and only those sensitive words with scores exceeding this threshold will be retained as target sensitive words. This score threshold may be determined based on historical advertising effect data, user behavior analysis, or expert experience. The purpose of filtering is to separate the true from the false from the initial sensitive words, and to eliminate those sensitive words with low frequency of occurrence, weak relevance, or little impact on user decision-making. In this way, only those highly relevant sensitive words that can truly represent user preferences and interests will be retained. Through this filtering process, a set of streamlined but high-quality target sensitive words can be obtained, providing more accurate input for subsequent advertising content optimization. These target sensitive words will be directly used to adjust the advertising language style, highlight product features, or strengthen emotional appeals, so that the advertising content is more in line with the actual needs and preferences of users, thereby improving the personalization, attractiveness, and ultimate delivery effect of the advertisement.

[0096] As an example, after the personalized advertisement is pushed and displayed and the advertisement delivery step is completed, the method further includes: obtaining user feedback; and adjusting the preset sensitive word weight according to the user feedback.

[0097] User feedback refers to the various responses and interactions of users to the personalized advertising content that has been pushed and displayed. These feedbacks can be collected in a variety of ways, including but not limited to: Click-Through Rate (CTR): The ratio of the number of times users click on an ad to the number of times the ad is displayed, which is an important indicator to measure the degree to which an ad attracts user attention; Conversion Rate: The ratio of the number of times users complete specific target actions (such as purchases, registrations, downloads, etc.) after clicking on an ad to the total number of clicks, reflecting the promotional effect of the ad; Interactive Behavior: The interaction between users and advertising content, such as likes, comments, and sharing, which indicate the user's attitude and engagement with the advertising content; Watching time: For video ads, the length of time users watch ads can reflect the attractiveness of the advertising content; Exit rate: The proportion of users who leave the page or app after watching an ad. A high exit rate may indicate that the advertising content is not attractive enough; User evaluation: Users' direct evaluation of the advertising content, which may be collected through questionnaires, comment areas, or customer service channels.

[0098] After personalized advertising is completed, feedback is collected by tracking users' actual behavior towards the advertisements. Specific practices include monitoring the number of clicks on the advertisements, recording the number of users who have completed purchases or other target actions, and analyzing the interactions between users and advertisements, such as comments and sharing. The purpose of this is to measure the actual effect of the advertisements and understand the real reactions of users to the advertisement content. Based on these feedbacks, the weight of sensitive words in the advertisements is adjusted. For example, if an advertisement with a sensitive word attracts a large number of user clicks and interactions, the system will increase the weight of this word so that it can be used more frequently in future advertisements, thereby increasing the attractiveness and conversion rate of the advertisements. On the contrary, if the advertisement of a certain word does not perform well, its weight will be reduced. In this way, the advertising strategy can be continuously optimized to make the advertising content more in line with the actual interests and preferences of users, and ultimately achieve the purpose of improving the advertising effect.

[0099] In this embodiment, the multimodal agent extracts the initial sensitive words from the user interaction data according to the sensitive word classification; the sensitive word score is calculated according to the preset first-level classification frequency, the preset first-level classification weight, the preset second-level classification frequency, the preset second-level classification weight, the initial sensitive words and the preset sensitive word weight; the initial sensitive words are filtered according to the sensitive word score to obtain the target sensitive words. First, the system analyzes the user interaction data through the multimodal agent and extracts the initial sensitive words according to the preset sensitive word classification system. This step can efficiently identify keywords related to user behavior and preferences from a large amount of data, providing rich materials for the personalization of advertising content. Then, the system calculates a comprehensive score based on the frequency of occurrence of sensitive words in user data, the preset classification weight and the weight of the sensitive words themselves. This score quantifies the importance of each sensitive word and provides an objective evaluation standard for advertising content optimization. Then, the system sets a score threshold to filter out those sensitive words with lower scores, and retains the sensitive words that exceed the threshold as target sensitive words. This filtering process ensures that advertising content optimization can focus on those sensitive words that are most likely to affect user decisions. Through this series of steps, the system can obtain a concise and highly relevant set of sensitive words. These sensitive words will be used to optimize the advertising content, making it more in line with the actual needs and preferences of users, improving the relevance and attractiveness of advertisements, and ultimately improving advertising effectiveness and user experience.

[0100] For example, to help understand the implementation process of the personalized advertising method obtained by combining this embodiment with the above-mentioned embodiment 1, please refer to Figure 3 , Figure 3 A brief flowchart of a personalized advertising delivery method is provided, specifically:

[0101] First, the real-time interactive content is input into the multimodal processing unit, which is responsible for extracting sensitive words and storing them in the database, while classifying and lexicalizing the sensitive words. The frequency of sensitive words in the database is used by the sensitive word filter to determine the importance of each sensitive word, that is, the sensitive word score. When a user enters a web page, the advertising system is triggered to pull the topK (first k) sensitive words related to the user from the database, and input these sensitive words into the advertising display optimizer together with the brand advertising content. The advertising display optimizer uses large model technology to combine the user's sensitive words and the original advertising content to optimize and generate new advertising display content to improve the personalization and attractiveness of the advertisement. Finally, the optimized advertising content is pushed to the user, and the system also adjusts the weight of the sensitive words based on the user's feedback on the advertisement, so as to achieve more accurate personalized matching in future advertising delivery. This process ensures that the advertising content can dynamically adapt to the needs and preferences of users, thereby improving the effectiveness of advertising and the user's experience satisfaction.

[0102] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the personalized advertising delivery method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0103] This application also provides a personalized advertising device, please refer to Figure 4 , the personalized advertising delivery device comprises:

[0104] The data acquisition module 10 is used to acquire user interaction data, initial advertisement content and sensitive word classification;

[0105] A data processing module 20, configured to process the user interaction data according to the sensitive word classification through a multimodal agent to obtain a target sensitive word;

[0106] An advertisement optimization module 30, configured to optimize and adjust the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content;

[0107] The advertisement push module 40 is used to push and display the target advertisement content to complete the advertisement delivery.

[0108] In one embodiment, the data processing module 20 is further used to extract initial sensitive words from the user interaction data according to the sensitive word classification through a multimodal intelligent agent; calculate sensitive word scores according to a preset first-level classification frequency, a preset first-level classification weight, a preset second-level classification frequency, a preset second-level classification weight, the initial sensitive words and a preset sensitive word weight; and filter the initial sensitive words according to the sensitive word scores to obtain target sensitive words.

[0109] In one embodiment, the advertisement optimization module 30 is further used to construct optimized prompt words according to the target sensitive words, the advertisement prompt words and the initial advertisement content; and the multimodal intelligent agent performs semantic analysis and content optimization on the initial advertisement content according to the optimized prompt words and the target sensitive words to obtain the target advertisement content.

[0110] In one embodiment, the advertising optimization module 30 is further used to perform semantic analysis on the initial advertising content through the multimodal intelligent agent to obtain original semantics; and integrate the target sensitive words into the initial advertising content according to the optimization prompt words and the original semantics through the multimodal intelligent agent to obtain target advertising content.

[0111] In one embodiment, the data processing module 20 is further used to obtain user feedback; and adjust the preset sensitive word weight according to the user feedback.

[0112] In one embodiment, the advertising push module 40 is further used to determine audience characteristics and audience preferences based on the user interaction data and the target sensitive words; determine the target advertising platform and advertising push parameters based on the audience characteristics and the audience preferences; obtain the format requirements and review requirements of the target advertising platform; adjust the target advertising content according to the format requirements and the review requirements to obtain the advertising content to be pushed; push and display the advertising content to be pushed according to the advertising push parameters to complete the advertising delivery.

[0113] In one embodiment, the data acquisition module 10 is also used to collect user interaction data through a preset platform; obtain user characteristics based on the user interaction data; obtain initial advertising content from a preset advertising database based on the user characteristics; and obtain sensitive word classification based on market research, user research and product feature definitions.

[0114] The personalized advertising delivery device provided by the present application adopts the personalized advertising delivery method in the above-mentioned embodiment, which can solve the technical problem of how to improve the effectiveness of advertising and the user's experience satisfaction. Compared with the prior art, the beneficial effects of the personalized advertising delivery device provided by the present application are the same as the beneficial effects of the personalized advertising delivery method provided by the above-mentioned embodiment, and other technical features in the personalized advertising delivery device are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0115] The present application provides a personalized advertising delivery device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the personalized advertising delivery method in the above-mentioned embodiment 1.

[0116] Reference below Figure 5 , which shows a schematic diagram of the structure of a personalized advertising delivery device suitable for implementing the embodiment of the present application. The personalized advertising delivery device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The personalized advertising delivery device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0117] like Figure 5 As shown, the personalized advertising delivery device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the personalized advertising delivery device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the personalized advertising delivery device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a personalized advertising delivery device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0118] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0119] The personalized advertising delivery device provided by the present application adopts the personalized advertising delivery method in the above embodiment, which can solve the technical problem of how to improve the effectiveness of advertising and the user's experience satisfaction. Compared with the prior art, the beneficial effects of the personalized advertising delivery device provided by the present application are the same as the beneficial effects of the personalized advertising delivery method provided by the above embodiment, and the other technical features in the personalized advertising delivery device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0120] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0121] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0122] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the personalized advertising delivery method in the above-mentioned embodiment.

[0123] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0124] The computer-readable storage medium may be included in the personalized advertisement delivery device; or may exist independently without being assembled into the personalized advertisement delivery device.

[0125] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the personalized advertising delivery device, the personalized advertising delivery device: obtains user interaction data, initial advertising content and sensitive word classification; processes the user interaction data according to the sensitive word classification through a multimodal intelligent agent to obtain target sensitive words; optimizes and adjusts the initial advertising content according to the multimodal intelligent agent and the target sensitive words to obtain target advertising content; pushes and displays the target advertising content to complete advertising delivery.

[0126] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0127] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0128] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0129] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned personalized advertising delivery method, and can solve the technical problem of how to improve the effectiveness of advertisements and the user's experience satisfaction. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the personalized advertising delivery method provided in the above-mentioned embodiment, and will not be repeated here.

[0130] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned personalized advertising delivery method when executed by a processor.

[0131] The computer program product provided in this application can solve the technical problem of how to improve the effectiveness of advertisements and the user's experience satisfaction. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the personalized advertisement delivery method provided in the above embodiment, which will not be repeated here.

[0132] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A personalized advertising delivery method, characterized in that: The method comprises: Obtain user interaction data, initial advertising content, and sensitive word classification; Processing the user interaction data according to the sensitive word classification by a multimodal agent to obtain target sensitive words; Optimizing and adjusting the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content; The target advertisement content is pushed and displayed to complete the advertisement delivery.

2. The method according to claim 1, characterized in that The step of processing the user interaction data according to the sensitive word classification by the multimodal agent to obtain the target sensitive word comprises: Extracting initial sensitive words from the user interaction data according to the sensitive word classification by a multimodal agent; The sensitive word score is calculated according to the preset first-level classification frequency, the preset first-level classification weight, the preset second-level classification frequency, the preset second-level classification weight, the initial sensitive word and the preset sensitive word weight; The initial sensitive words are filtered according to the sensitive word scores to obtain target sensitive words.

3. The method according to claim 1, characterized in that The multimodal agent is constructed based on the multimodal large model and the advertising prompt words, and the step of optimizing and adjusting the initial advertising content based on the multimodal agent and the target sensitive words to obtain the target advertising content includes: Constructing an optimized prompt word according to the target sensitive word, the advertising prompt word and the initial advertising content; The multimodal intelligent agent performs semantic analysis and content optimization on the initial advertisement content according to the optimization prompt words and the target sensitive words to obtain target advertisement content.

4. The method according to claim 3, characterized in that The step of performing semantic analysis and content optimization on the initial advertisement content according to the optimization prompt words and the target sensitive words by the multimodal agent to obtain the target advertisement content comprises: Performing semantic analysis on the initial advertisement content by the multimodal agent to obtain original semantics; The multimodal agent integrates the target sensitive words into the initial advertisement content according to the optimized prompt words and the original semantics to obtain the target advertisement content.

5. The method according to claim 2, characterized in that After the step of pushing and displaying the personalized advertisement and completing the step of advertisement delivery, the method further includes: Get user feedback; The preset sensitive word weight is adjusted according to the user feedback.

6. The method according to claim 1, characterized in that The step of pushing and displaying the target advertisement content to complete advertisement delivery includes: Determining audience characteristics and audience preferences based on the user interaction data and the target sensitive words; Determine a target advertising platform and advertising push parameters according to the audience characteristics and the audience preferences; Obtaining the format requirements and review requirements of the target advertising platform; Adjust the target advertisement content according to the format requirement and the review requirement to obtain the advertisement content to be pushed; The advertisement content to be pushed is pushed and displayed according to the advertisement push parameters to complete the advertisement delivery.

7. The method according to any one of claims 1 to 6, characterized in that The step of obtaining user interaction data, initial advertisement content, and sensitive word classification includes: Collect user interaction data through the preset platform; Obtaining user features according to the user interaction data; Acquire initial advertisement content from a preset advertisement database according to the user characteristics; Sensitive words are classified based on market research, user surveys and product feature definitions.

8. A personalized advertising device, characterized in that: The device comprises: Data acquisition module, used to obtain user interaction data, initial advertising content and sensitive word classification; A data processing module, configured to process the user interaction data according to the sensitive word classification through a multimodal agent to obtain target sensitive words; An advertisement optimization module, configured to optimize and adjust the initial advertisement content according to the multimodal agent and the target sensitive word to obtain target advertisement content; The advertisement push module is used to push and display the target advertisement content to complete the advertisement delivery.

9. A personalized advertising device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the personalized advertisement delivery method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the personalized advertising delivery method according to any one of claims 1 to 7 are implemented.

Images