Evaluation system for live broadcast advertisement

Abstract

The invention discloses a live broadcast advertisement evaluation system. The system comprises a data acquisition module, an emotion analysis module, an advertisement effect evaluation module, a real-time feedback module and a data storage and analysis module. The data acquisition module is used for acquiring interaction data, advertisement playing data and emotion feedback data of a user in real time. The sentiment analysis module analyzes user comment sentiment tendency through a natural language processing technology and calculates sentiment scores. The advertisement effect evaluation module comprehensively evaluates the advertisement effect through a formula E based on the user interaction degree, the advertisement exposure degree, the emotion score and the conversion rate, and weight parameters are dynamically optimized through a machine learning algorithm. The real-time feedback module feeds back the advertisement effect to an advertiser in real time in a chart form, and the data storage and analysis module performs regression analysis, clustering analysis and trend prediction through historical data to provide support for advertisement strategy optimization.

Description

Technical Field

[0001] The present invention relates to the technical field of advertisement evaluation, and in particular to an evaluation system for live broadcast advertisements. Background Art

[0002] With the rise of live e-commerce, social live broadcasting and various online live broadcasting platforms, live broadcasting advertising has become an important means for businesses to promote their products and services. Through real-time interaction between anchors and users, advertising can be more naturally integrated into live broadcast content, thereby attracting audience attention, increasing brand exposure and product conversion rates. Compared with traditional TV ads, search engine ads and social media ads, live broadcasting ads have the advantages of real-time interactivity and precise delivery, which can achieve efficient matching of advertising information with potential consumers.

[0003] However, with the rapid development of live broadcast advertising, how to scientifically and accurately evaluate the advertising effect has become an important issue that needs to be solved. Existing technologies mainly focus on traditional advertising evaluation methods and simple data statistics, which have the following shortcomings:

[0004] Traditional advertising evaluation technology mainly evaluates advertising effectiveness through post-event statistics and audience feedback surveys. For example, advertising performance is evaluated through advertising exposure, click-through rate, number of views, and sales data. This method has the following problems:

[0005] Poor timeliness: Statistics can only be collected after the advertisement is released, and it is impossible to monitor the advertising effect in real time during the advertisement playback process.

[0006] Long feedback cycle: Traditional audience feedback surveys take time, and the survey results often lag behind the actual advertising, making it difficult to reflect the advertising effect in a timely manner.

[0007] Lack of real-time optimization capabilities: Due to the lack of real-time data feedback, advertisers are unable to adjust advertising content or delivery strategies in a timely manner and may miss the best promotion opportunities.

[0008] In the live broadcast scenario, user interaction is real-time and the advertising effect changes dynamically with the progress of the live broadcast. However, existing evaluation technologies make it difficult to monitor and evaluate the advertising effect in real time and comprehensively during the live broadcast, resulting in advertisers being unable to grasp the advertising performance in a timely manner.

[0009] Existing advertising effectiveness evaluation technologies mainly rely on single or limited data indicators, such as exposure, click-through rate and number of views. Although these indicators can reflect the basic performance of advertisements, they cannot fully reflect the actual impact of advertisements on viewers. Specific problems include:

[0010] Ignoring user interaction data: User interaction behaviors such as likes, comments, and shares are important indicators for measuring audience interest and ad popularity, but existing technologies do not pay enough attention to this.

[0011] Lack of sentiment analysis: Users’ emotional reactions (positive, negative, or neutral) when watching advertisements directly affect the acceptance of advertisements and brand image. However, existing technologies lack the ability to analyze the sentiment tendency of user comments.

[0012] Lack of multi-dimensional integration: Existing technologies often analyze a certain indicator in isolation and fail to integrate user interaction data, exposure data, emotional feedback and conversion rate in multiple dimensions, resulting in one-sided and inaccurate advertising effectiveness evaluation results.

[0013] The ultimate goal of live broadcast advertising is user conversion, that is, to encourage users to make purchases or other targeted actions after watching the advertisement. However, the existing technology has the following deficiencies in conversion rate evaluation:

[0014] Unable to establish the relationship between exposure and conversion: Although it is possible to count ad exposure, it is difficult to track the entire process from user viewing the ad to actual conversion.

[0015] Lack of behavioral data correlation: Failure to deeply analyze the relationship between user behavior paths and interactive behaviors and conversion rates leads to inaccurate conversion rate evaluation.

[0016] Ignoring intermediate factors: Conversion behavior is affected by multiple factors, such as user emotions, degree of interaction, and ad playback duration, but existing technologies fail to comprehensively consider these factors, resulting in low accuracy in conversion rate evaluation.

[0017] The weight parameters in traditional advertising evaluation models (such as exposure, interaction, and emotional feedback) are often fixed. These parameters are usually set based on experience and cannot be dynamically optimized based on actual advertising scenarios and user feedback. Specific problems include:

[0018] Unable to adapt to different advertising scenarios: factors such as the advertising target audience, delivery time, and advertising format have different impacts on advertising effectiveness, but fixed weights cannot adapt to these changes.

[0019] Lack of adaptive capabilities: The existing system lacks intelligent optimization capabilities based on historical data and is unable to dynamically adjust weight parameters by learning from historical advertising performance, resulting in inaccurate evaluation results.

[0020] Most existing technologies are limited to data collection and display, and lack a closed-loop optimization mechanism based on historical data analysis and feedback, which is specifically reflected in:

[0021] Data isolation: Advertising effectiveness data is only used as a single evaluation indicator, and it is unable to analyze and predict trends through historical data to provide optimization suggestions.

[0022] Lack of closed-loop feedback: It is impossible to form a closed loop from data collection, effect evaluation to strategy adjustment, which makes it difficult for advertisers to optimize advertising strategies in a timely manner based on evaluation results, thereby reducing the return on investment (ROI) of advertising.

[0023] In summary, existing advertising effect evaluation technologies have the following problems: insufficient real-time performance, single analysis dimension, inaccurate conversion rate evaluation, rigid evaluation model, and lack of closed-loop feedback mechanism. These deficiencies are particularly prominent in live broadcast advertising scenarios, because live broadcast advertising has the characteristics of strong real-time performance, rich interactive data, and fast emotional feedback. Traditional advertising evaluation technologies cannot meet the needs of efficient and accurate evaluation of live broadcast advertising.

[0024] Therefore, we urgently need to design a live broadcast advertising evaluation system to solve the above problems. Summary of the invention

[0025] The purpose of the present invention is to provide an evaluation system for live broadcast advertisements in order to solve the technical problems raised in the above-mentioned background technology.

[0026] The technical solution adopted by the present invention to achieve the above-mentioned purpose is:

[0027] The solution of the present invention provides a live broadcast advertisement evaluation system, which includes:

[0028] The data collection module is used to collect user behavior data, advertisement playback data and emotional data in real time during the live broadcast process. The data includes: 观看 :The number of users who viewed the ad;

[0029] N 点赞： Number of likes by users; N 评论 :Number of user comments; n 分享： Number of times users share; t 观看 :Total ad viewing time; T 总投放 :Total duration of advertisement delivery; N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative emotions in user comments; N 转化 :The number of users converted through advertising;

[0030] Sentiment analysis module: used to analyze user comments based on natural language processing technology and extract sentiment score S;

[0031] Advertising effect evaluation module: used to calculate the comprehensive score E of advertising effect based on the collected data, where:

[0032] e=α·I+β·P+γ·S+θ·C;

[0033] Among them: α, β, γ, θ are weight parameters, satisfying α+β+γ+θ=1; I: user interaction; P: advertising exposure; S: sentiment score; C: conversion rate;

[0034] Real-time feedback module: used to provide real-time feedback of advertising effect evaluation results to advertisers in the form of charts;

[0035] Data storage and analysis module: used to store collected data and historical advertising performance data, and provide analysis and optimization suggestions.

[0036] As a preferred technical solution of the present invention, the user interaction degree I is calculated by the following formula:

[0037]

[0038] Where: N 观看 : The number of users who watched the ad; n 点赞： Number of likes by users; n 评论 : Number of user comments; N 分享： Number of times a user has shared.

[0039] As a preferred technical solution of the present invention, the advertising exposure p is calculated by the following formula:

[0040]

[0041] Where: T 观看 :Total ad viewing time, T 总投放 :Total duration of advertisement delivery.

[0042] As a preferred technical solution of the present invention, the sentiment score S is calculated by the following formula:

[0043]

[0044] Where: N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative sentiments in user comments.

[0045] As a preferred technical solution of the present invention, the conversion rate C is calculated by the following formula:

[0046]

[0047] Where: N 转化 N: The number of users converted through advertising; 观看 :The number of users who viewed the ad.

[0048] As a preferred technical solution of the present invention, the weight parameters α, β, γ, θ are dynamically adjusted through a machine learning algorithm, specifically including: setting the optimization goal to maximize the comprehensive score E of the advertising effect;

[0049] Train optimal weight parameters based on historical advertising data through models such as regression analysis, gradient boosting or neural networks;

[0050] Output the weight combination adjusted in real time to meet the effect requirements set by advertisers.

[0051] As a preferred technical solution of the present invention, the data acquisition module realizes data acquisition through the live broadcast platform API interface or embedded code, and supports real-time data transmission.

[0052] As a preferred technical solution of the present invention, the real-time feedback module displays the results as a bar chart, line chart or pie chart through a data visualization tool to display the advertising interactivity, exposure, sentiment score and conversion rate.

[0053] As a preferred technical solution of the present invention, the data storage and analysis module analyzes the historical advertising data by the following method:

[0054] Regression analysis: evaluate the contribution of user interaction behavior to advertising effectiveness;

[0055] Cluster analysis: Identify the differences in effectiveness in different time periods and different advertising formats;

[0056] Trend analysis: Extract the time trend of emotional feedback and conversion rate to predict future advertising effects.

[0057] The present invention also provides a live broadcast advertisement evaluation method, the method comprising the following steps:

[0058] Data collection: Real-time collection of user interaction data, advertising playback data, and emotional data during live broadcasts;

[0059] Effect calculation: The advertising effect score is calculated based on the formula E = α·I + β·P + γ·S + θ·C; where: α, β, γ, θ are weight parameters, satisfying α+β+γ+θ=1; I: user interaction; P: advertising exposure; S: sentiment score; C: conversion rate; S: sentiment score; C: conversion rate;

[0060] Real-time feedback: Provide real-time feedback of effect evaluation results to advertisers;

[0061] Historical analysis and optimization: Optimize weight parameters through machine learning algorithms based on historical data;

[0062] Strategy adjustment and re-delivery: Advertisers adjust their advertising strategies based on feedback results to complete the effect optimization closed loop.

[0063] Compared with the prior art, the present invention provides the following beneficial effects:

[0064] 1. The present invention uses a data acquisition module to obtain user interaction data, advertising playback data and emotional feedback data in real time during the live broadcast process, thereby achieving real-time monitoring of advertising effects. At the same time, the system integrates multi-dimensional indicators such as user interaction, advertising exposure, emotional score and conversion rate to provide a comprehensive advertising effect evaluation, overcoming the defects of the traditional advertising evaluation that is lagging and single.

[0065] 2. The present invention introduces a machine learning algorithm to dynamically optimize the weight parameters of the advertising effect evaluation model based on historical data, so that the evaluation results are more in line with the actual scenario, and the accuracy and intelligence of the evaluation are improved. In addition, the advertising effect evaluation module calculates E through a multi-dimensional quantitative formula to achieve accurate quantification of advertising performance, helping advertisers quickly locate problems and optimize delivery strategies.

[0066] 3. The present invention forms a closed-loop system of data collection, effect evaluation, real-time feedback and strategy optimization through two-way interaction between the data storage and analysis module and other modules. The system can not only display the advertising effect in real time, but also provide optimization suggestions through historical data analysis, helping advertisers to improve the return on investment (ROI) of advertising, and achieve scientific decision-making and continuous optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 A system block diagram of a live broadcast advertisement evaluation system;

[0069] Figure 2 The present invention is a flow chart of a method for evaluating a live broadcast advertisement. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] The following is combined with Figure 1 , Figure 2With reference to the accompanying drawings and a number of embodiments, specific embodiments of the present invention are described in detail.

[0072] Implementation method 1: system architecture and module implementation;

[0073] The present invention provides a live broadcast advertisement evaluation system, the overall architecture of which includes:

[0074] Data acquisition module;

[0075] Sentiment analysis module;

[0076] Advertising effectiveness evaluation module;

[0077] Real-time feedback module;

[0078] Data storage and analysis module;

[0079] Data acquisition module;

[0080] This module collects the following data in real time by embedding the API interface or front-end code of the live broadcast platform:

[0081] User behavior data: N 观看 : The number of users who watched the ad; N 点赞： Number of likes by users; N 评论 : Number of user comments; N 分享： Number of times users share;

[0082] Advertisement playback data: T 观看 :Total ad viewing time; T 总投放 :Total duration of advertisement delivery;

[0083] User sentiment data: N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative emotions in user comments;

[0084] Conversion data: N 转化 :The number of users converted by the ad.

[0085] Sentiment analysis module: uses natural language processing (NLP) technology to analyze the sentiment tendency of user comments:

[0086] Use sentiment lexicon or pre-trained sentiment classification model to annotate the review text;

[0087] Categorize reviews as positive, neutral, or negative based on their content;

[0088] Count positive sentiment data N 正面 and negative sentiment data N 负面 .

[0089] Calculate the sentiment score S: Where: N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative sentiments in user comments.

[0090] Advertising effect evaluation module: The advertising effect score is calculated according to the following formula: E = α·I + β·P + γ·S + θ·C, where: α, β, γ, θ are weight parameters, satisfying α+β+γ+θ=1; I: user interaction; P: advertising exposure; S: sentiment score; C: conversion rate.

[0091] in: N 观看 : The number of users who watched the ad; N 点赞： Number of likes by users; N 评论 : Number of user comments; N 分享： Number of times users share;

[0092] P: Advertising exposure: Where: T 观看 :Total ad viewing time, T 总投放 :Total duration of advertisement delivery;

[0093] S: Sentiment score: Where: N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative emotions in user comments;

[0094] Conversion rate C: Where: N 转化 N: The number of users converted through advertising; 观看 :The number of users who viewed the ad.

[0095] Real-time feedback module: Use data visualization tools to generate bar charts, line charts and other charts to display advertising performance indicators, including interaction, exposure, sentiment score and conversion rate. Real-time update of performance scores makes it easier for advertisers to monitor advertising performance.

[0096] Data storage and analysis module: stores real-time collected data and advertising effect scores;

[0097] Through historical data analysis, regression analysis, cluster analysis and trend analysis are carried out to provide support for advertising strategy optimization.

[0098] Implementation method 2: weight optimization based on machine learning;

[0099] Weight parameter optimization: The weight parameters α, β, γ, and θ are dynamically adjusted through machine learning algorithms, including: Optimization goal: maximize the advertising effect score E.

[0100] Training method: Based on historical advertising data, use gradient boosting tree (GBDT) or neural network model to model advertising performance and output the optimal weight combination.

[0101] Feedback mechanism: Dynamically adjust the weight according to the real-time advertising effect score E to improve the accuracy of advertising evaluation.

[0102] The following is a further detailed description in conjunction with several embodiments:

[0103] Example 1: Evaluation of a brand’s advertising on a live streaming platform Live streaming duration: 60 minutes Advertisement playback data:

[0104] T 观看 =1800;

[0105] T 总投放 =3600.

[0106] User interaction data:

[0107] N 观看 =500 people;

[0108] N 点赞 =300 times;

[0109] N 评论 =200;

[0110] N 分享 =50 times.

[0111] User sentiment data:

[0112] N 正面 =160;

[0113] N 负面 =40.

[0114] Conversion data:

[0115] N 转化 =20 people.

[0116] Calculation results: Interaction I: Exposure P:

[0117]

[0118] Sentiment score S:

[0119]

[0120] Conversion rate C:

[0121]

[0122] Advertising effect score E (assuming weights α = 0.4, β = 0.2, γ = 0.3, θ = 0.1):

[0123] E=0.4·1.1+0.2·0.5+0.3·0.6+0.1·0.04=0.44+0.1+

[0124] 0.18+0.004=0.7;

[0125] Result analysis: The ad score is 0.724, the emotional feedback is good, but the conversion rate is low. The system feedback suggests optimizing the ad content to increase the user conversion rate.

[0126] Example 2: Comparative analysis of advertising formats:

[0127] Compare two advertising formats for the same brand:

[0128] Ad A: Video ad, high engagement, positive sentiment score;

[0129] Ad B: Image ad, with high exposure but low engagement.

[0130] Through system analysis:

[0131] Ad A: E=0.8, with a greater contribution to interaction.

[0132] Advertisement B: E=0.6, exposure is the main contributing factor.

[0133] Optimization suggestions: Based on the advertiser's goals, adjust the advertising format appropriately to balance exposure and interaction effects.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A live broadcast advertisement evaluation system, characterized in that: include: The data collection module is used to collect user behavior data, advertisement playback data and emotional data in real time during the live broadcast process. The data includes: 观看 : The number of users who watched the ad; N 点赞： Number of likes by users; N 评论 : Number of user comments; N 分享： Number of times users share; T 观看 :Total ad viewing time; T 总投放 :Total duration of advertisement delivery; N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative emotions in user comments; N 转化 :The number of users converted through advertising; Sentiment analysis module: used to analyze user comments based on natural language processing technology and extract sentiment score S; Advertising effect evaluation module: used to calculate the comprehensive score E of advertising effect based on the collected data, where: E=α·I+β·P+γ·S+θ·C; Among them: α, β, γ, θ are weight parameters, satisfying α+β+γ+θ=1; I: user interaction; P: advertising exposure; S: sentiment score; C: conversion rate; Real-time feedback module: used to provide real-time feedback of advertising effect evaluation results to advertisers in the form of charts; Data storage and analysis module: used to store collected data and historical advertising performance data, and provide analysis and optimization suggestions.

2. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The user interaction degree i is calculated by the following formula: Where: N 观看 : The number of users who watched the ad; N 点赞： Number of likes by users; N 评论 : Number of user comments; N 分享： Number of times a user has shared.

3. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The advertising exposure P is calculated by the following formula: Where: T 观看 :Total ad viewing time, T 总投放 :Total duration of advertisement delivery.

4. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The sentiment score S is calculated by the following formula: Where: N 正面 : The number of positive emotions in user comments; N 负面 : The number of negative sentiments in user comments.

5. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The conversion rate C is calculated by the following formula: Where: N 转化 N: The number of users converted through advertising; 观看 :The number of users who viewed the ad.

6. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The weight parameters α, β, γ, and θ are dynamically adjusted through a machine learning algorithm, specifically including: setting the optimization goal to maximize the comprehensive score E of the advertising effect; Train optimal weight parameters based on historical advertising data through models such as regression analysis, gradient boosting or neural networks; Output the weight combination adjusted in real time to meet the effect requirements set by advertisers.

7. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The data collection module realizes data collection through the live broadcast platform API interface or embedded code, and supports real-time data transmission.

8. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The real-time feedback module displays the results as a bar chart, line chart or pie chart through a data visualization tool, showing the advertising interaction, exposure, sentiment score and conversion rate.

9. The live broadcast advertisement evaluation system according to claim 1, characterized in that: The data storage and analysis module analyzes historical advertising data by the following method: Regression analysis: evaluate the contribution of user interaction behavior to advertising effectiveness; Cluster analysis: Identify the differences in effectiveness in different time periods and different advertising formats; Trend analysis: Extract the time trend of emotional feedback and conversion rate to predict future advertising effects.

10. An evaluation method based on the live broadcast advertisement evaluation system according to any one of claims 1 to 9, characterized in that: The method comprises the following steps: Data collection: Real-time collection of user interaction data, advertising playback data, and emotional data during live broadcasts; Effect calculation: The advertising effect score is calculated based on the formula E = α·I + β·P + γ·S + θ·C; where: α, β, γ, θ are weight parameters, satisfying α+β+γ+θ=1; I: user interaction; P: advertising exposure; S: sentiment score; C: conversion rate; S: sentiment score; C: conversion rate; Real-time feedback: Provide real-time feedback of effect evaluation results to advertisers; Historical analysis and optimization: Optimize weight parameters through machine learning algorithms based on historical data; Strategy adjustment and re-delivery: Advertisers adjust their advertising strategies based on feedback results to complete the effect optimization closed loop.

Images