Streaming media advertisement intelligent putting method and system

Abstract

The invention discloses an intelligent streaming media advertisement putting method and system, and particularly relates to the technical field of big data analysis, social media interaction data and watching behavior data of a user are collected in real time through a background of a streaming media platform, the collected data are analyzed in real time, a deep learning algorithm is used, the collected multi-dimensional data are combined, and the intelligent streaming media advertisement putting method and system are obtained. The method comprises the following steps: establishing a personalized interest model of a user, predicting a most relevant advertisement type, automatically screening out advertisement contents most relevant to the current interest, emotion and behavior of the user from an advertisement library according to a real-time data analysis result, carrying out dynamic adjustment, and carrying out advertisement insertion through an advertisement insertion technology of a streaming media platform. The advertisement content is seamlessly embedded into the video playing process, the advertisement insertion mode is adjusted according to the real-time user watching state, in the advertisement playing process, an advertisement effect report is generated according to the interactive feedback and watching behavior of the user, and future advertisement putting is optimized and adjusted through data analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data analysis, more particularly, the present application relates to a streaming media advertisement intelligent delivery method and system. BACKGROUND

[0002] With the popularity and development of the Internet, streaming media advertising, as a new form of advertising delivery, has become an important part of the advertising industry. Especially in smart TVs, mobile devices and various streaming media platforms, the way and strategy of advertising delivery are constantly evolving, showing a trend of increasing intelligence, personalization and dynamization.

[0003] In existing advertising delivery technology, the user's historical behavior information is mainly relied on to generate interest tags, and advertising is pushed based on this. Although this method uses the user's past behavior to speculate their interests, it does not take into account the user's real-time interest changes. The user's interest may change over time, context or other external factors, and only through historical behavior tags may not accurately capture the user's current interest state, resulting in insufficient timeliness and relevance of advertising delivery. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a streaming media advertisement intelligent delivery method and system to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme, a streaming media advertisement intelligent delivery method, specifically comprising: Through the background of the streaming media platform, real-time collection of user social media interaction data and viewing behavior data, and real-time analysis of the collected data, including: viewing content analysis, emotional feedback analysis and social media analysis; Using deep learning algorithms, combined with the collected multi-dimensional data, a personalized interest model of the user is established, and the most relevant advertisement type is predicted; According to the real-time data analysis result, automatically filter out the most relevant advertisement content from the advertisement library according to the user's current interest, emotion and behavior, and dynamically adjust according to the three standards of advertisement form selection, advertisement creative customization and advertisement adaptation; Through the advertisement insertion technology of the streaming media platform, the advertisement content is seamlessly embedded into the video playback process, the insertion mode of the advertisement is adjusted according to the real-time user viewing state, and the advertisement is designed as user-interactive content; During the advertising playback process, an advertising effect report is generated according to the user's interactive feedback and viewing behavior, and the future advertising delivery is optimized and adjusted through data analysis.

[0006] Preferably, as a preferred scheme of the intelligent streaming media advertising method, the method comprises collecting user social media interaction data and viewing behavior data in real time through the background of the streaming media platform, and performing real-time analysis on the collected data, including viewing content analysis, emotional feedback analysis and social media analysis, specifically including: The API interface of the social media platform is used to obtain the interaction data of the user, the player API and the background system of the streaming media platform are used to obtain the viewing behavior data of the user, embedded sensors and data transmission rate monitoring means are used to obtain data of device performance and network conditions in real time, and the collected raw data is cleaned, standardized, missing value filled and denoised, and then stored in a database; The viewing behavior data includes currently viewed video content, video playback progress, video type, viewing time period and user stay duration, and the change of user interest is dynamically analyzed based on the viewing behavior data of the user by using big data analysis and machine learning algorithms, including viewing content analysis, emotional feedback analysis and social media analysis.

[0007] Preferably, as a preferred scheme of the intelligent streaming media advertising method, the viewing content analysis is based on interaction data and viewing behavior data to analyze user interest and preference, the interaction data includes the number of likes, the number of comments and the number of shares, and the social interaction index is calculated based on the interaction data as , wherein L is the number of likes, C is the number of comments, and S is the number of shares, is the weight coefficient of each index, and the social interaction index is used to calculate the user interest score , the content is sorted according to the user interest, and related video content is pushed; The emotional feedback analysis is an emotional classification of the text in the user comments and social media interactions by natural language processing, and an emotional score is generated for each interaction as , represents negative emotion, represents neutral emotion, represents positive emotion, the emotional score of the user in different time periods is calculated, and the emotional change of the user in the viewing process is obtained , wherein is the total number of interactions at time point t, is the total number of interactions at time point t-1, is the emotional score of the i th interaction at time point t, is the emotional score of the i th interaction at time point t-1, is the emotional change in the time period t; The social media analysis is to evaluate the propagation effect of the video content according to the interaction of the user on the social platform, and to combine the social interaction index with the propagation effect to obtain a comprehensive propagation index CPI, and the specific formula is Wherein, L is the number of likes, C is the number of comments, S is the sharing amount, F is the number of fans of the video publisher, And are weight coefficients, and SI is a social interaction index.

[0008] Preferably, as a preferred scheme of the flow media advertisement intelligent putting method, it includes using a deep learning algorithm, combining the collected multi-dimensional data to establish a personalized interest model of the user, and predicting the most relevant advertisement type, which specifically includes: The multi-dimensional data collected is converted into numerical features by One-Hot encoding, including user behavior features, interaction features, and emotional feedback features, and a convolutional neural network is used to extract image and visual features in the video content, which are combined into a feature vector , which is used as input into a deep neural network model, wherein, is the image and visual feature vector, is the emotional feedback feature, is the user behavior feature after One-Hot encoding, is the emotional feedback feature; A deep neural network model is established to predict the user's interest in different advertisement types, and the output is an interest score of each advertisement type, and the advertisement type is recommended according to the highest score.

[0009] Preferably, as a preferred scheme of the flow media advertisement intelligent putting method, it includes establishing a deep neural network model to predict the user's interest in different advertisement types, and the output is an interest score of each advertisement type, and the advertisement type is recommended according to the highest score, and further includes: A multi-layer perceptron structure is constructed, including an input layer, a hidden layer and an output layer, the feature input of the input layer is the combined user features, and the size is k+m, wherein k is the feature dimension based on One-Hot encoding, and m is the visual feature dimension extracted; The hidden layer is a plurality of fully connected layers, and the output of each layer is transformed by a nonlinear activation function: , wherein, is the activation function, is the output of the i-th layer, And are the weight and bias items of the layer, is the input layer; The output layer is an advertisement interest score vector , which represents the user's interest level in each ad type, and the output vector is: where W is the weight of the fully connected layer, h is the input of the current layer, b is the bias term, and f is the activation function. Each is the user's interest score for ad type M; The model is trained using mean square error (MSE) as the loss function, which calculates the error between the user's interest prediction and the true label. The true interest label is represented as y, and the loss function is: where is the predicted interest score, is the true interest label, and N is the number of samples; The Adam optimizer is used to minimize the loss function and update the weights in the network. Based on the trained model, the user's feature vector is used to predict their interest score p for each ad type, and the ad type with the highest interest score is selected where is the ad type that the user is most interested in. Based on the predicted interest score, the timing and type of ad insertion are dynamically adjusted to improve the click-through rate and user engagement.

[0010] Preferably, as a preferred solution of the intelligent streaming media ad placement method, it includes automatically filtering the most relevant ad content to the user's current interest, emotion and behavior from the ad library based on real-time data analysis results, and dynamically adjusting according to three standards of ad form selection, ad creative customization and ad adaptation, specifically including: Using natural language processing and computer vision techniques to analyze the current video content, identifying ad-related elements, and dynamically adjusting the timing and format of ad display based on the user's current behavior and emotion, the specific formula is: where represents the display time of the ad, and are weights representing the influence of user behavior and ad effectiveness, and represent the user's social interaction index and comprehensive propagation index at time t, respectively; Based on the creative and adaptive features of the ad content, the form of the ad is dynamically adjusted by optimizing the objective function, which is: where is the type and form of the ad content, is the weight of the ad creative effect, measuring the attractiveness of the ad creative and the user emotion matching degree, is the weight of the ad effect, reflecting the historical effect of the ad, is the weight of the adaptability of the advertisement, indicating the adaptability of the advertisement in different platforms, devices and network environments.

[0011] Preferably, as a preferred scheme of the intelligent streaming media advertisement delivery method, it includes the advertisement insertion technology through the streaming media platform, seamlessly embedding the advertisement content into the video playing process, adjusting the insertion mode of the advertisement according to the real-time user watching state, and designing the advertisement as user-participable content, specifically including: adopting the forms of question and answer, investigation and lottery to encourage users to participate in the design of the advertisement content, adjusting the display duration and insertion mode of the advertisement according to the user watching state, including static, short video and interactive advertisement, and making the advertisement naturally integrated through content preheating and gradual effect means, and the specific formula of the advertisement display form is: , wherein, represents the display form of the advertisement, including picture, video and interactive advertisement, is the continuity score of the advertisement display, that is, how the advertisement is seamlessly combined with the video content to avoid interrupting the user experience, is the display duration adjustment of the advertisement, flexibly adjusting the advertisement duration according to the user watching state, and are weights, respectively reflecting the influence of the continuity and duration optimization of the advertisement display on the advertisement effect; After the advertisement is displayed, user behavior data is collected, including clicking, interacting and skipping, and the advertisement display mode is optimized and adjusted through real-time data feedback: , wherein, represents the update of the advertisement display strategy, is the click rate of the advertisement, measuring the attractiveness of the advertisement, is the cost per interaction, measuring the cost performance ratio of the advertisement investment and effect, and are the weights of the click rate and interaction cost, reflecting the optimization direction of the advertisement effect.

[0012] Preferably, as a preferred scheme of the intelligent streaming media advertisement delivery method, it includes the generation of the advertisement effect report according to the user interaction feedback and watching behavior during the advertisement playing process, the optimization and adjustment of future advertisement delivery through the analysis of the data, specifically including: According to the user interaction behavior and the data after the advertisement display, the platform generates an advertisement effect report, including: the change trend of the click rate and the interaction rate, the click rate and the advertisement investment and output ratio, the ratio of the user skipping the advertisement, and the adaptability of the advertisement; Based on the advertising effect report, specific optimization adjustment suggestions are proposed, including adjusting the advertising creativity, optimizing the advertising duration and frequency, strengthening the advertising personalization and targeted delivery, and budget optimization and resource allocation, further including: The adjustment of the advertising creativity is to optimize the advertising copy, visual design, and interactive mode according to the changes in click-through rate and interaction rate, to improve the attractiveness and participation of the advertisement; The optimization of the advertising duration and frequency is to reduce the advertising duration and frequency for the advertisements with high skip rate, to avoid disturbing the user experience; The strengthening of the advertising personalization and targeted delivery is to further strengthen the personalized recommendation and precise targeted delivery of the advertisement according to the adaptability analysis of the advertisement, to improve the matching degree of the advertisement and the audience; The budget optimization and resource allocation is to adjust the advertising budget allocation based on the input-output ratio, to invest more budget in efficient advertising strategies, and to reduce the investment in inefficient advertisements.

[0013] The application also provides a streaming media advertising intelligent delivery system, specifically including a user data acquisition and analysis module, a personalized interest model construction module, an advertising screening and dynamic adjustment module, an advertising insertion and interactive design module, and an advertising effect monitoring and optimization module; The user data acquisition and analysis module is to collect the social media interaction data and viewing behavior data of the user in real time through the background of the streaming media platform, and to analyze the collected data in real time, including viewing content analysis, emotional feedback analysis, and social media analysis; The personalized interest model construction module is to use deep learning algorithm to establish the personalized interest model of the user based on the collected multi-dimensional data, and to predict the most relevant advertising type; The advertising screening and dynamic adjustment module is to automatically screen the most relevant advertising content from the advertising library according to the real-time data analysis results, and to dynamically adjust according to the three standards of advertising form selection, advertising creativity customization, and advertising adaptation; The advertising insertion and interactive design module is to seamlessly embed the advertising content into the video playback process through the advertising insertion technology of the streaming media platform, to adjust the insertion mode of the advertisement according to the real-time user viewing state, and to design the advertisement as user-interactive content; The advertising effect monitoring and optimization module generates an advertising effect report according to the user's interactive feedback and viewing behavior during the advertising playback process, and optimizes the future advertising delivery through data analysis.

[0014] In another aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements the steps of the method for intelligent streaming media advertisement placement as described above.

[0015] In another aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for intelligent streaming media advertisement placement as described above.

[0016] In the above technical solution, the present application has the following advantages: By combining more real-time data, including the user's current viewing content, emotional feedback, social media interaction to dynamically adjust the advertisement placement, the user's behavior and emotional changes can be analyzed in real time, the most relevant advertisement content can be accurately pushed, the relevance of the advertisement and the user participation can be improved, and through the integration of the streaming media platform, the advertisement can be seamlessly inserted into the video content, the advertisement interference can be reduced, the viewing rate and user acceptance of the advertisement can be improved through the fusion with the content. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0018] Figure 1 The method flowchart of the present application.

[0019] Table 1 is a data recording table of the simulation experiment of the present application. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] In the description of the present application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0022] In the description of the present application, the term "for example" is used to mean "serving as an example, instance, or illustration." Any embodiment described as "for example" in this application is not necessarily to be construed as preferred or advantageous over other embodiments. The following description is presented to enable any person skilled in the art to make and use the application. In the following description, for purposes of explanation, specific details are set forth. It will be apparent to those skilled in the art that the present application can be practiced without the specific details. In other instances, well-known structures and processes are not shown in detail to avoid obscuring the application. Thus, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features presented herein.

[0023] Embodiment 1 The present embodiment provides a streaming media advertising intelligent delivery method as shown, specifically comprising: Figure 1 Through the background of the streaming media platform, real-time collection of user social media interaction data and viewing behavior data, and real-time analysis of the collected data, including: viewing content analysis, emotional feedback analysis and social media analysis; Using deep learning algorithms, combined with the collected multi-dimensional data, a personalized interest model of the user is established, and the most relevant advertising type is predicted; According to the real-time data analysis results, automatically filter out the most relevant advertising content from the advertising library according to the user's current interest, emotion and behavior, and dynamically adjust according to the three standards of advertising form selection, advertising creative customization and advertising adaptation; Through the advertising insertion technology of the streaming media platform, the advertising content is seamlessly embedded into the video playback process, and the insertion method of the advertising is adjusted according to the real-time user viewing state, and the advertising is designed as user-interactive content; During the advertising playback process, according to the user's interactive feedback and viewing behavior, an advertising effect report is generated, and through the analysis of the data, the future advertising delivery is optimized and adjusted.

[0024] Preferably, the real-time collection of user social media interaction data and viewing behavior data through the background of the streaming media platform, and the real-time analysis of the collected data, including: viewing content analysis, emotional feedback analysis and social media analysis, specifically comprising: Using the API interface of the social media platform to obtain the user's interaction data, through the player API and background system of the streaming media platform to obtain the user's viewing behavior data, using embedded sensors and data transmission rate monitoring means to obtain real-time data of device performance and network status, cleaning, standardizing, missing value filling and denoising the collected raw data, and storing it in the database;​ The viewing behavior data includes currently viewed video content, video playback progress, video type, viewing period, and user dwell time. Based on the user viewing behavior data, the dynamic analysis of user interest changes is performed using big data analysis and machine learning algorithms, including viewing content analysis, emotional feedback analysis, and social media analysis, which further includes: The viewing content analysis is based on interactive data and viewing behavior data to analyze user interests and preferences. The interactive data includes the number of likes, the number of comments, and the number of shares. Based on the interactive data, the social interaction index is calculated as where L is the number of likes, C is the number of comments, and S is the number of shares. is the weight coefficient of each indicator. Based on the calculated social interaction index The content is sorted according to user interest and relevant video content is pushed. The emotional feedback analysis is performed by natural language processing on user comments and social media interaction text to generate an emotional score for each interaction as , represents negative emotions, represents neutral emotions, represents positive emotions, and the emotional score of the user in different time periods is calculated to obtain the emotional changes of the user during the viewing process where is the total number of interactions at time point t, is the total number of interactions at time point t−1, is the emotional score of the i-th interaction at time point t, is the emotional score of the i-th interaction at time point t−1, is the emotional change in time period t. The social media analysis is based on user interactions on social platforms to evaluate the propagation effect of video content and combine the social interaction index with the propagation effect to obtain the comprehensive propagation index CPI, with the specific formula being where L is the number of likes, C is the number of comments, S is the number of shares, and F is the number of fans of the video publisher. and are weight coefficients, and SI is the social interaction index.

[0025] Preferably, the deep learning algorithm is used to combine the collected multi-dimensional data to establish a personalized interest model of the user and predict the most relevant advertisement type, which specifically includes: The collected multi-dimensional data is converted into numerical features through One-Hot encoding, including user behavior features, interaction features, and emotional feedback features. Convolutional neural networks are used to extract image and visual features from video content, which are combined into a feature vector. , for input into a deep neural network model, wherein, is an image and a visual feature vector, is an emotional feedback feature, is a user behavior feature after One-Hot encoding, is an emotional feedback feature; A deep neural network model is established to predict the user's interest in different types of advertisements, and the output is the interest score of each advertisement type, and the advertisement type with the highest score is recommended, further comprising: A multi-layer perceptron structure is constructed, including an input layer, a hidden layer and an output layer, the feature input of the input layer is the combined user feature, with a size of k+m, where k is the feature dimension based on One-Hot encoding, and m is the visual feature dimension extracted; The hidden layer is a plurality of fully connected layers, and the output of each layer is transformed by a nonlinear activation function: wherein, is an activation function, is the output of the i-th layer, and are the weights and bias terms of the layer, is the input layer; The output layer is an advertisement interest score vector , indicating the degree of interest of the user to each type of advertisement, and for M types of advertisements, the output vector is: where W is the weight of the fully connected layer, h is the input of the current layer, b is the bias term, and f is the activation function. Each is the user's interest score for the advertisement type M; The mean square error MSE is used as the loss function to train the model, and the error between the user's interest prediction and the true label is calculated, and the true interest label is represented as y, and the loss function is: wherein, is the predicted interest score, is the true interest label, and N is the number of samples; The Adam optimizer is used to minimize the loss function to update the weights in the network, and based on the user's feature vector , the interest score p of each type of advertisement is predicted, and the advertisement type with the highest interest score is selected, is the advertisement type that the user is most interested in, based on the predicted interest score, the timing and type of advertisement insertion are dynamically adjusted to improve the click-through rate and user engagement of the advertisement.

[0026] Preferably, the most relevant advertisement content to the user's current interest, emotion and behavior is automatically selected from the advertisement library based on real-time data analysis results, and dynamically adjusted according to three standards of advertisement form selection, advertisement creative customization and advertisement adaptation, which specifically includes: The current video content is analyzed using natural language processing and computer vision technology, the elements related to the advertisement are identified, and the timing and format of the advertisement display are dynamically adjusted based on the user's current behavior and emotion, and the specific formula is: , wherein, represents the display time of the advertisement, and are weights, respectively representing the influence of user behavior and advertisement effect, and respectively represent the user social interaction index and the comprehensive propagation index at time t; Based on the creative and adaptive characteristics of the advertisement content, the form of the advertisement is dynamically adjusted by optimizing the objective function, and the objective function is: , wherein, is the type and form of the advertisement content, is the weight of the advertisement creative effect, which measures the attractiveness of the advertisement creative and the matching degree of the user emotion, is the weight of the advertisement effect, which reflects the historical effect of the advertisement, is the weight of the advertisement adaptation, which represents the adaptability of the advertisement in different platforms, devices and network environments.

[0027] Preferably, the advertisement content is seamlessly embedded into the video playback process through the advertisement insertion technology of the streaming media platform, the insertion mode of the advertisement is adjusted according to the real-time user viewing state, and the advertisement is designed as user-interactive content, which specifically includes: Adopting the forms of question and answer, survey and lottery, encouraging users to participate in the design of advertisement content, adjusting the display time and insertion mode of the advertisement according to the user viewing state, including static, short video, interactive advertisement, and making the advertisement naturally integrated through content preheating and gradient effect means, and the specific formula of the advertisement display form is: , wherein, represents the display form of the advertisement, including picture, video, interactive advertisement, is the continuity score of the advertisement display, that is, how the advertisement is seamlessly combined with the video content to avoid interrupting the user experience, is the display time adjustment of the advertisement, which flexibly adjusts the advertisement time according to the user's viewing state, and are weights, respectively reflecting the influence of the continuity and time optimization of the advertisement display on the advertisement effect; After the advertisement is displayed, user behavior data including clicking, interaction, skipping are collected, and the advertisement display mode is optimized and adjusted through real-time data feedback: wherein, represents the update of the advertisement display strategy, is the click rate of the advertisement, which measures the attractiveness of the advertisement, is the cost per interaction, which measures the cost-effectiveness of the advertisement investment and effect, and are the weights of the click rate and the interaction cost, reflecting the optimization direction of the advertisement effect.

[0028] Preferably, during the playing of the advertisement, an advertisement effect report is generated according to the user's interaction feedback and viewing behavior, and the future advertisement investment is optimized and adjusted through the analysis of the data, specifically including: According to the user's interaction behavior and the data after the advertisement is displayed, the platform generates an advertisement effect report, including: the change trend of the click rate and the interaction rate, the click rate and the advertisement investment and output ratio, the ratio of users skipping the advertisement, and the adaptability of the advertisement; Based on the advertisement effect report, specific optimization and adjustment suggestions are proposed, including adjusting the advertisement creativity, optimizing the advertisement duration and frequency, strengthening the advertisement personalization and targeted delivery, and budget optimization and resource allocation, further including: The adjustment of the advertisement creativity is to optimize the advertisement copy, visual design and interaction mode according to the change of the click rate and the interaction rate, to improve the attractiveness and participation of the advertisement; The optimization of the advertisement duration and frequency is to reduce the advertisement duration and frequency for the advertisement with high skipping rate, to avoid disturbing the user experience; The strengthening of the advertisement personalization and targeted delivery is to further strengthen the personalized recommendation and accurate targeted delivery of the advertisement according to the adaptability analysis of the advertisement, to improve the matching degree of the advertisement and the audience; The budget optimization and resource allocation is to adjust the advertisement budget allocation based on the input-output ratio, to invest more budget in efficient advertisement strategies, and to reduce the investment in inefficient advertisements.

[0029] Embodiment 2 The embodiment provides a streaming media advertisement intelligent delivery system, specifically including a user data acquisition and analysis module, a personalized interest model construction module, an advertisement screening and dynamic adjustment module, an advertisement insertion and interaction design module, and an advertisement effect monitoring and optimization module; The user data acquisition and analysis module is to collect the social media interaction data and viewing behavior data of the user in real time through the background of the streaming media platform, and to analyze the collected data in real time, including: viewing content analysis, emotional feedback analysis and social media analysis; The personalized interest model construction module is used to establish a personalized interest model of a user and predict the most relevant advertisement type by using a deep learning algorithm and combining with collected multi-dimensional data. The advertisement screening and dynamic adjustment module is used to automatically screen the most relevant advertisement content from an advertisement library according to real-time data analysis results, and dynamically adjust the advertisement content according to three standards of advertisement form selection, advertisement creative customization and advertisement adaptation. The advertisement insertion and interaction design module is used to seamlessly embed advertisement content into a video playing process by using an advertisement insertion technology of a streaming media platform, adjust the insertion mode of the advertisement according to real-time user watching state, and design the advertisement as content that can be participated by the user. The advertisement effect monitoring and optimization module is used to generate an advertisement effect report according to interactive feedback and watching behavior of the user during advertisement playing, and optimize and adjust future advertisement delivery by analyzing data.

[0030] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to perform wired or wireless communication with an external terminal. The wireless communication can be achieved by WIFI, a carrier network, NFC (near field communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, a trackball or a touchpad arranged on the shell of the computer device. In addition, the input device can also be an external keyboard, a touchpad or a mouse, etc.

[0031] The embodiment also provides a storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps of the method for intelligently delivering streaming media advertisements according to the above embodiment; and the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk, or an optical disk.

[0032] Embodiment 3 The following is another embodiment of the present application, which provides a method for intelligently delivering streaming media advertisements. In order to verify the beneficial effects of the present application, a simulation experiment is performed for scientific demonstration.

[0033] The experiment aims to verify the effectiveness of the method for intelligently delivering streaming media advertisements. Through data collection, analysis, dynamic adjustment of advertisement content, and collection of user feedback, the relevance, viewing rate, and user engagement of the advertisements are improved. The experiment collects user behavior data and emotional feedback in real time, and optimizes the delivery of advertisements by combining deep learning algorithms. The simulation experiment is performed according to the method for intelligently delivering streaming media advertisements provided in Embodiment 1. The specific steps include, Through the streaming media platform background, user social media interaction data and viewing data are collected in real time. Embedded sensors and device feedback are used to monitor device performance and network conditions in real time. The data collection frequency is set to once per minute to ensure real-time data. Deep learning algorithms are used to analyze the data in real time, including user viewing content, emotional feedback, and social media interaction. Based on the above multi-dimensional data, a personalized interest model of the user is established, and the most relevant advertisement type to the user's current emotions and behaviors is predicted. The advertisement content is dynamically adjusted according to the user's real-time emotions and behaviors. Based on the real-time data analysis results, the most relevant advertisement content is selected from the advertisement library. The advertisement content includes advertisement form, advertisement creative customization, and advertisement adaptation. The advertisement content is dynamically adjusted to improve the viewing rate of the advertisement and the engagement of the user. Using the advertisement insertion technology of the streaming media platform, the advertisement is seamlessly embedded into the video playing process, and the advertisement content is designed as interactive content that users can participate in; During the advertisement playing process, an advertisement effect report is generated according to the interactive feedback and watching behavior of the user, and through analyzing the advertisement effect report, the future advertisement delivery strategy is optimized.

[0034] The specific data of the above simulation experiment is as follows: Time / minute User viewing content User emotional feedback Social media interaction User behavior characteristics Advertisement content type Advertisement effect prediction User interaction feedback Advertisement insertion strategy Advertisement effect report 0-5 Action movie Excitement, joy Like Long viewing time Game advertisement High relevance High engagement Interrupted segment insertion advertisement High click rate, high engagement 5-10 Comedy movie Relaxation, happiness Share Medium viewing time Beauty product advertisement Medium relevance Medium engagement Advertisement insertion by time node Lower click rate 10-15 Documentary Thinking, immersion Comment Preference for documentary Travel advertisement High relevance High engagement Dynamic advertisement insertion High engagement 15-20 Science fiction movie Excitement, surprise Like, share High interaction Technology product advertisement Extremely high relevance Extremely high engagement Advertisement engagement design Excellent user feedback 20-25 Animation movie Happiness, joy Share, comment Long viewing time Children's toy advertisement Medium relevance High engagement Advertisement insertion by viewing time High engagement 25-30 Documentary Contemplation, calmness Like, comment Short pause Environmentally friendly product advertisement High relevance Medium engagement Advertisement adjustment by emotional feedback Medium click rate Table 1 Experimental analysis: By comparing the coincidence degree of the advertisement effect data and the user behavior, the accuracy of the system in advertisement accurate delivery and user emotion prediction is verified, and the experimental results show that through the deep learning algorithm and the personalized recommendation model, the most relevant advertisement content can be accurately predicted, and the advertisement content is dynamically adjusted according to the user emotional feedback and social interaction, which significantly improves the watching rate and participation rate of the advertisement; when the advertisement content is dynamically adjusted according to the user's interest, emotion and behavior, the user's advertisement participation rate is significantly improved, and in the form of interactive advertisement and customized advertisement, the user is more willing to participate and has a higher click rate.

[0035] The above only describes certain exemplary embodiments of the present application in a descriptive manner, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

Claims

1. A method for intelligent delivery of streaming media advertisements, characterized in that: Specifically, it includes: Through the backend of the streaming media platform, real-time data on users' social media interactions and viewing behavior is collected, and the collected data is analyzed in real time, including: content viewing analysis, sentiment feedback analysis, and social media analysis. Using deep learning algorithms and combining collected multi-dimensional data, a personalized interest model of users is built, and the most relevant ad types are predicted. Based on real-time data analysis results, the system automatically selects the most relevant advertising content to the user's current interests, emotions, and behaviors from the advertising library, and dynamically adjusts the content according to three criteria: advertising format selection, advertising creative customization, and advertising adaptation. By using the ad insertion technology of streaming media platforms, advertising content can be seamlessly embedded into the video playback process. The insertion method of the ads can be adjusted according to the real-time viewing status of users, and the ads can be designed to be interactive content for users. During ad playback, an ad performance report is generated based on user interaction feedback and viewing behavior. By analyzing the data, future ad placements can be optimized and adjusted.

2. The intelligent delivery method for streaming media advertising according to claim 1, characterized in that: The system collects users' social media interaction data and viewing behavior data in real time through the backend of the streaming media platform, and performs real-time analysis on the collected data, including: viewing content analysis, sentiment feedback analysis, and social media analysis, specifically including: The system uses the API interfaces of social media platforms to obtain user interaction data, the player API and backend system of streaming media platforms to obtain user viewing behavior data, and uses embedded sensors and data transmission rate monitoring methods to obtain device performance and network status data in real time. The collected raw data is cleaned, standardized, missing value imputation and noise reduction are performed, and then stored in the database. The viewing behavior data includes the currently viewed video content, video playback progress, video type, viewing time period, and user dwell time. Using big data analysis and machine learning algorithms, based on the user's viewing behavior data, changes in user interests are dynamically analyzed, including: viewing content analysis, sentiment feedback analysis, and social media analysis.

3. The intelligent delivery method for streaming media advertising according to claim 2, characterized in that: The dynamic analysis of changes in user interests based on user viewing behavior data includes: viewing content analysis, sentiment feedback analysis, and social media analysis, and further includes: The content viewing analysis is based on interaction data and viewing behavior data to analyze user interests and preferences. The interaction data includes the number of likes, comments, and shares. A social interaction index is calculated based on this interaction data. Where L is the number of likes, C is the number of comments, and S is the number of shares. The weighting coefficients for each indicator are based on the calculated social interaction index. The content is sorted according to user interests, and relevant video content is pushed to them. The sentiment feedback analysis uses natural language processing to classify the sentiment of text in user comments and social media interactions, generating a sentiment score for each interaction. , Indicates negative emotions. Indicates a neutral mood. Representing positive emotions, the system calculates users' emotional scores at different time points to reveal changes in their emotions during the viewing process. ,in, It is the total number of interactions at time point t. It is the total number of interactions at time point t−1. It is the emotion score of the i-th interaction at time point t. It is the sentiment score of the i-th interaction at time point t−1. It refers to the emotional changes within a time period t; The aforementioned social media analysis assesses the dissemination effect of video content based on user interactions on social platforms, and combines the social interaction index with the dissemination effect to obtain the comprehensive dissemination index (CPI). The specific formula is as follows: Where L represents the number of likes, C represents the number of comments, S represents the number of shares, and F represents the number of followers of the video uploader. and It is the weighting coefficient, and SI is the social interaction index.

4. The intelligent delivery method for streaming media advertising according to claim 1, characterized in that: The process involves using deep learning algorithms, combined with collected multi-dimensional data, to build a personalized interest model for users and predict the most relevant ad types. Specifically, this includes: One-Hot encoding is used to transform the collected multi-dimensional data into numerical features, including user behavior features, interaction features, and emotional feedback features. Convolutional neural networks are then used to extract image and visual features from the video content and combine them into a feature vector. This is used as input to the deep neural network model, where... These are image and visual feature vectors. It is an emotional feedback characteristic. These are user behavior features encoded in One-Hot format. It is a characteristic of emotional feedback; Build a deep neural network model to predict users' interest in different ad types. The output is an interest score for each ad type, and the model recommends ad types based on the highest score.

5. The intelligent delivery method for streaming media advertising according to claim 4, characterized in that: The step of establishing a deep neural network model to predict users' interest in different ad types, with the output being an interest score for each ad type, and recommending ad types based on the highest score, further includes: Construct a multilayer perceptron structure, including an input layer, a hidden layer and an output layer. The feature input of the input layer is the combined user features of size k+m, where k is the feature dimension based on One-Hot encoding and m is the dimension of the extracted visual features. The hidden layer consists of multiple fully connected layers, and the output of each layer is transformed by a non-linear activation function: ,in, It is an activation function. It is the output of the i-th layer. and These are the weights and biases of this layer. For input layer; The output layer is an advertising interest score vector. , representing the user's level of interest in each type of advertisement. For M types of advertisements, the output vector is: Where W is the weight of the fully connected layer, h is the input of the current layer, b is the bias term, and f is the activation function. It is the user's interest rating for ad type M; The model is trained using the mean squared error (MSE) as the loss function. The error between the predicted user interest in the advertisement and the actual label is calculated. The actual interest label is denoted as y, and the loss function is: ,in, It is a predicted interest rating. These are real interest tags, and N is the number of samples; The Adam optimizer is used to minimize the loss function, and the weights in the network are updated based on the trained model and the user's feature vector. Predict their interest score p for each ad type, and select the ad type with the highest interest score. ,in, It identifies the types of ads that users are most interested in, and dynamically adjusts the timing and type of ad insertion based on predicted interest scores.

6. The intelligent delivery method for streaming media advertising according to claim 1, characterized in that: Based on real-time data analysis results, the system automatically filters advertising content from the ad library that is most relevant to the user's current interests, emotions, and behaviors, and dynamically adjusts it according to three criteria: ad format selection, ad creative customization, and ad adaptation. Specifically, this includes: The system uses natural language processing and computer vision techniques to analyze the current video content, identify elements relevant to the advertisement, and dynamically adjust the timing and format of the advertisement display based on the user's current behavior and emotions. The specific formula is as follows: ,in, Indicates the moment when the advertisement is displayed. and These are weights, representing the impact of user behavior and advertising effectiveness, respectively. and These represent the user social interaction index and the comprehensive communication index at time t, respectively. Based on the creativity and suitability characteristics of the advertising content, the form of the advertisement is dynamically adjusted by optimizing the objective function, which is: ,in, It refers to the type and format of the advertising content. It is the weight of the advertising creative effect. It is the weight of advertising effectiveness. It is the weight of ad relevance.

7. The intelligent delivery method for streaming media advertising according to claim 1, characterized in that: The aforementioned advertising insertion technology using streaming media platforms seamlessly embeds advertising content into the video playback process, adjusts the insertion method based on real-time user viewing status, and designs the advertisements as user-engageable content, specifically including: The ad format employs Q&A, surveys, and prize draws to encourage user participation in ad content design. The ad display duration and insertion method are adjusted based on user viewing behavior. Content pre-heating and gradient effects are used to seamlessly integrate the ad. The specific formula for the ad display format is as follows: ,in, This refers to the display format of the advertisement, including images, videos, and interactive ads. It is the score for the consistency of the ad display. It's an adjustment to the display duration of the advertisement. and These are weights, which respectively reflect the impact of ad display continuity and duration optimization on ad performance; After the ad is displayed, user behavior data is collected, including clicks, interactions, and skips. The ad display method is optimized and adjusted based on real-time data feedback. ,in, This indicates an update to the ad display strategy. It's the click-through rate of an ad, a measure of its attractiveness. It's the cost of each interaction. and It is the weight of click-through rate and cost of interaction.

8. The intelligent delivery method for streaming media advertising according to claim 1, characterized in that: During the ad playback process, an ad performance report is generated based on user interaction feedback and viewing behavior. Through data analysis, future ad placements are optimized and adjusted, specifically including: Based on user interaction and post-ad display data, the platform generates an ad performance report, including: Trends in click-through rate and engagement rate, click-through rate and return on investment (ROI), user skip rate, and ad suitability; Based on the advertising performance report, specific optimization and adjustment suggestions are proposed, including adjusting advertising creatives, optimizing advertising duration and frequency, enhancing advertising personalization and targeted delivery, and optimizing budget and resource allocation.

9. A streaming media advertising intelligent delivery system applied to a streaming media advertising intelligent delivery method as described in any one of claims 1-8, characterized in that: It includes modules for user data collection and analysis, personalized interest model construction, ad filtering and dynamic adjustment, ad insertion and interaction design, and ad performance monitoring and optimization. The user data collection and analysis module collects users' social media interaction data and viewing behavior data in real time through the backend of the streaming media platform, and performs real-time analysis on the collected data, including: viewing content analysis, sentiment feedback analysis, and social media analysis. The personalized interest model building module uses deep learning algorithms, combined with collected multi-dimensional data, to build a personalized interest model for users and predict the most relevant ad types. The ad filtering and dynamic adjustment module automatically filters ad content most relevant to the user's current interests, emotions, and behaviors from the ad library based on real-time data analysis results, and dynamically adjusts it according to three criteria: ad format selection, ad creative customization, and ad adaptation. The ad insertion and interaction design module uses the ad insertion technology of the streaming media platform to seamlessly embed ad content into the video playback process, adjust the ad insertion method according to the real-time user viewing status, and design the ad into content that users can participate in. During ad playback, the ad performance monitoring and optimization module generates an ad performance report based on user interaction feedback and viewing behavior. Through data analysis, it optimizes and adjusts future ad placements.

Images