Advertisement generation system based on creative characteristics

Abstract

The invention relates to the technical field of data processing, in particular to a creative feature-based advertisement generation system, which comprises an acquisition module, an analysis module, an evaluation module, a decision module and a regulation and control module, the acquisition module is used for acquiring information data parameters generated by advertisements of creative characteristics of a target platform and copywriting information data parameters of a creative model in a historical period; the analysis module is used for analyzing the information data feature value and the copywriting information data feature value of the creative model; the evaluation module is used for judging whether advertisement generation meets a standard; the decision-making module is used for judging whether the creative model generated by the advertisement which does not meet the standard meets the standard or not; and the regulation and control module is used for adjusting a preset information data characteristic threshold value and giving an alarm. According to the method, the matching degree and the accuracy of the creative characteristic technical indexes and the market demand indexes in creative advertisement generation are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to an advertisement generation system based on creative features. BACKGROUND

[0002] In the current field of intelligent creative advertisement generation, generative models trained based on historical data are generally used for the automatic production of advertisement content. However, in the existing technical solutions, there is a serious gap and lag between the internal technical indicators of creative features and the external demand indicators of market effects. In the model optimization process of the existing system, static internal technical parameters such as copy error rate, semantic similarity, and keyword matching degree are mainly used as the basis for evaluation and updating. Although these indicators can guarantee the basic quality and standardization of creative content, they cannot truly and timely reflect the dynamically changing market preferences and user feedback. When the optimization signal cannot directly anchor the ultimate market utility of click-through rate and praise rate, the iteration direction of the creative model deviates from the market demand, resulting in a serious delay in model updating. The outdated model cannot generate advertisement content that fits the current user interest and competitive environment, which directly leads to a continuous decline in advertisement click-through rate and praise rate, and ultimately results in a continuous decrease in the accuracy and effectiveness of the entire creative advertisement generation system.

[0003] Chinese Patent Publication No. CN120563168A discloses a multi-language advertisement creative generation system, which includes a data collection and preprocessing module responsible for collecting and preprocessing cultural and language data. A cultural feature analysis and modeling module extracts cultural features from the data and constructs a cultural model to map the features and language expressions. A creative generation module constructs a semantic understanding model based on deep learning and interfaces with the cultural model to generate creative themes and plots. A language generation module inputs the creative themes and target language into the language generation model to output advertisement scripts. A multi-language optimization and integration module optimizes the scripts in different languages and integrates them into an advertisement creative set. A user interaction and feedback module provides an interface for inputting cultural tourism projects, target languages, and parameters, outputs scripts, and optimizes the system according to user feedback. The present application can improve the efficiency of cultural tourism advertisement creative generation, enhance the adaptability of advertisement scripts in different language and cultural environments, and break through the limitations of traditional advertisement creative generation.

[0004] CN119477427A discloses an advertisement planning method and system based on artificial intelligence, which comprises: obtaining advertisement feature data; preliminarily processing the obtained advertisement feature data to obtain a preliminary advertisement feature data set; dividing audience groups based on the preliminary advertisement feature data set using a K-means algorithm; extracting keyword sequences based on the audience groups; generating advertisement copy sequences using a generative adversarial network model based on the keyword sequences; calculating similarity and fluency based on the advertisement copy sequences, obtaining a comprehensive score by weighted average, and selecting the advertisement copy with the highest comprehensive score as the final output. This method can effectively improve the relevance and appeal of the advertisement copy, optimize the advertisement delivery effect, ultimately improve the click-through rate and conversion rate of users, bring higher investment returns to advertisers, and greatly improve the creativity and delivery effect of advertisements.

[0005] Therefore, the prior art has the following problems: In the creative advertisement generation process, there is a difference between the creative feature technical index and the market demand index, which leads to the fact that the creative model is not updated in time, resulting in a decrease in advertisement click-through rate and good comment rate, and further leading to low precision of creative advertisement generation. SUMMARY

[0006] Therefore, the present application provides an advertisement generation system based on creative features to overcome the problem of low precision of creative advertisement generation caused by the difference between the creative feature technical index and the market demand index in the creative advertisement generation process of the prior art.

[0007] To achieve the above purpose, the present application provides an advertisement generation system based on creative features, which comprises: A collection module is used to collect information data parameters of creative feature advertisement generation and copy information data parameters of creative models in a historical period; An analysis module is connected to the collection module and is used to analyze information data feature values based on information data parameters; and analyze copy information data feature values based on copy information data parameters of creative models; An evaluation module is connected to the analysis module and is used to determine whether the advertisement generation meets the standard based on the information data feature values; If the generated advertisement does not meet the standard, the copy information data parameters of the creative model of the advertisement generation that does not meet the standard are continuously collected; the copy information data feature values are analyzed based on the copy information data parameters; and it is determined whether the creative model of the advertisement generation meets the standard based on the difference between the copy information data feature values and the predetermined copy information data feature threshold value; a decision module connected with the analysis module and the evaluation module, configured to determine whether the creative model of the non-standard advertisement generation conforms to the standard based on the characteristic value of the copy information data; a regulation module connected with the decision module, configured to generate a corresponding processing strategy based on the non-standard advertisement generation; if the creative model of the non-standard advertisement generation conforms to the standard, it is determined that the creative model is effective, and it is determined to adjust the predetermined information data characteristic threshold value; if the creative model of the non-standard advertisement generation does not conform to the standard, it is determined that the creative model is invalid, and it is determined that the regulation module issues an alarm; wherein, the information data parameters include: click rate and praise rate; the creative model copy data parameters include: the matching degree of copy and keyword and the length of copy.

[0008] Further, the collection module is used to collect the information data parameters of the creative characteristics of the advertisement generation in the historical period of the target platform, which are click rate and praise rate; and the collection module is used to collect the copy information data parameters of the creative model, which are the matching degree of copy and keyword and the length of copy.

[0009] Further, the information data characteristic value analyzed by the analysis module is determined based on the sum of the first characteristic limiting representation parameter and the second characteristic limiting representation parameter, wherein, the first characteristic limiting representation parameter is the ratio of the click rate of the advertisement generation to the predetermined click rate; the second characteristic limiting representation parameter is the ratio of the praise rate of the advertisement generation to the predetermined praise rate.

[0010] Further, the copy information data characteristic value of the creative model analyzed by the analysis module is determined based on the sum of the first copy data limiting parameter and the second copy data limiting parameter, wherein, the first copy data limiting parameter is the ratio of the matching degree of copy and keyword to the predetermined matching degree of copy and keyword; the second copy data limiting parameter is the ratio of the predetermined length of copy to the length of copy.

[0011] Further, the evaluation module is used to determine the condition that the advertisement generation conforms to the standard, which is that the information data characteristic value is greater than the predetermined information data characteristic threshold value.

[0012] Further, the evaluation module is used to determine the condition that the advertisement generation does not conform to the standard, which is that the information data characteristic value is less than or equal to the predetermined information data characteristic threshold value.

[0013] Furthermore, the decision module determines that the creative model generated by the non-standard advertisement meets the standard if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is greater than a predetermined difference threshold.

[0014] Furthermore, the decision module determines that the creative model generated by the non-compliant advertisement does not meet the standard if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is less than or equal to a predetermined difference threshold.

[0015] Furthermore, the control module is used to adjust the predetermined information data feature threshold conditions so that the creative model generated by the advertisement that does not meet the standard meets the standard.

[0016] Furthermore, the condition for the control module to issue an alarm is that the creative model generated by the non-compliant advertisement does not meet the standards.

[0017] Compared with existing technologies, the beneficial effect of this invention is that it provides an advertising generation system based on creative features. By constructing a system comprising a data collection module, an analysis module, an evaluation module, a decision-making module, and a control module, this invention achieves advertising generation based on creative features. The data collection module systematically collects historical data parameters of ad generation and copywriting information from creative models, constructing a comprehensive, multi-dimensional data warehouse. This provides a solid data foundation for subsequent analysis, evaluation, and optimization, reducing the blindness of decisions based on experience or intuition. The analysis module performs in-depth mining and feature extraction on the collected raw data, identifying strong correlations between data features and the high performance of ad generation. This allows for objective measurement and comparison of ad and copy quality, laying the foundation for establishing scientific evaluation standards. The evaluation module, based on predetermined standards, performs rapid and consistent automated evaluation of the results of each batch or each ad generation, greatly improving review efficiency and reducing the cost and subjective bias of manual review. The decision-making module further diagnoses whether the creative model meets the standards when ad generation fails to meet them, determining whether it is a random fluctuation in a single generation or a defect in the creative model's capabilities. The implementation example reduces one-size-fits-all misjudgments through precise positioning, making optimization measures more targeted. The control module transforms the results of the analysis and decision-making modules into specific control actions, ultimately automating the entire process from problem discovery to problem resolution, achieving closed-loop intelligent management of the ad generation and optimization process.

[0018] In particular, the feature values ​​of the information data analyzed by the analysis module are determined based on the sum of a first feature-limited representation parameter and a second feature-limited representation parameter, wherein the first feature-limited representation parameter is the ratio of the click-through rate to the predetermined click-through rate; and the second feature-limited representation parameter is the ratio of the positive review rate to the predetermined positive review rate. The feature values ​​of the copywriting information data of the creative model analyzed by the analysis module are determined based on the sum of a first copywriting data-limited parameter and a second copywriting data-limited parameter, wherein the first copywriting data-limited parameter is the ratio of the matching degree between the copywriting and keywords to the predetermined matching degree between the copywriting and keywords; and the second copywriting data-limited parameter is the ratio of the predetermined copywriting length to the copywriting length. The comprehensive evaluation system, centered on the sum of multi-dimensional parameters and constrained representation values ​​of information data features and parameters of copywriting data, achieves an upgrade from single, one-sided judgments to comprehensive and stable evaluations, with significant beneficial effects. By comprehensively considering key factors such as click-through rate and positive review rate, copywriting relevance and standardization, the evaluation results are more comprehensive and objective, effectively reducing decision-making bias and providing precise data-driven support for resource optimization and creative iteration. The model transforms indicators of different dimensions into comparable coefficients through normalization, enhancing the system's anti-interference ability and stability, providing clear optimization objectives, and laying the foundation for intelligence. The implementation example powerfully supports operational automation and scalability through a quantifiable framework, while its flexible architecture facilitates future expansion to new evaluation dimensions, thereby achieving comprehensive improvements in accuracy, robustness, and efficiency.

[0019] In particular, by introducing a two-tiered evaluation and decision-making mechanism consisting of information data feature values ​​and copywriting information data feature values, the system achieves the beneficial effects of standardization and intelligentization of the advertising generation process. The system first uses multi-dimensional feature values ​​of comprehensive click-through rate and positive review rate to conduct preliminary screening of advertising generation. By comparing with predetermined thresholds, it efficiently distinguishes between advertisements that meet the standards and those that do not, ensuring basic quality. For advertisements that do not meet the standards in the preliminary screening, the system does not directly reject them, but initiates a secondary decision-making process. By evaluating the matching degree of its copywriting and keywords and the difference between its copywriting length and the predetermined copywriting, it intelligently judges whether there is any value in optimization. This not only establishes a stable and comprehensive automated evaluation benchmark, greatly improving the efficiency of the review, but more importantly, it has the dual ability to select the best and correct the worst. This ensures the overall quality of the generated advertisements and reduces the possibility of misjudging potential creative ideas due to poor performance of a single indicator, making the system's decision-making more refined and its resources more fully utilized.

[0020] In particular, by introducing two processing measures—dynamic adjustment of predetermined information data feature thresholds and alarm triggering implementations—the system's adaptability, resource utilization efficiency, and risk management level are significantly improved. These implementation measures continuously raise the system's quality entry threshold, driving the creative model to continuously optimize and surpass historical averages. This effectively reduces model stagnation, ensuring that ad generation performance adapts to market changes and continuously improves, achieving dynamic optimization of the system benchmark. The alarm mechanism triggered when the optimized model still does not meet the standards constitutes an important risk control and resource protection barrier. It promptly notifies operations personnel to manually intervene in substandard creatives or directly terminate ineffective optimization processes, preventing continuous waste of computing resources and operating costs, and effectively intercepting the delivery of low-quality ads, protecting user experience and brand image. These two measures complement each other: the former drives the system to continuously improve, while the latter ensures timely loss mitigation in operations, together forming a closed-loop intelligent management system that combines progressiveness and robustness. Attached Figure Description

[0021] Figure 1 This is a diagram illustrating the architecture of an advertising generation system based on creative features, as described in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the steps of generating an advertisement based on creative features according to an embodiment of the present invention. Figure 3 This is a logic diagram for determining whether an advertisement generation conforms to a standard in an embodiment of the present invention; Figure 4 This is a logic diagram for determining whether a creative model conforms to the standard in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] Please see Figure 1 The diagram shown is an architecture diagram of an advertising generation system based on creative features according to an embodiment of the present invention. The present invention provides an advertising generation system based on creative features, comprising: The data collection module is used to collect information data parameters of the creative characteristics of the target platform during the historical period, as well as copy information data parameters of the creative model. An analysis module, connected to the acquisition module, is used to analyze the feature values ​​of information data based on information data parameters; and to analyze the feature values ​​of copywriting information data based on the copywriting information data parameters of the creative model. An evaluation module, which is connected to the analysis module, is used to determine whether the generated advertisement meets the standards based on the feature values ​​of the information data; If the generated advertisement does not meet the standard, the text information data parameters of the creative model generated by the advertisement that does not meet the standard will continue to be collected; the text information data feature values ​​will be analyzed based on the text information data parameters; and the difference between the text information data feature values ​​and the predetermined text information data feature threshold will be used to determine whether the creative model generated by the advertisement meets the standard. The decision module, which is connected to the analysis module and the evaluation module, is used to determine whether the creative model generated by the non-standard advertisement meets the standard based on the feature values ​​of the copywriting information data. Control module: It is connected to the decision module and is used to propose corresponding processing strategies based on non-standard advertisement generation; If the creative model generated by the non-standard advertisement meets the standard, then the creative model is determined to be valid, and the predetermined information data feature threshold is determined to be adjusted. If the creative model generated by the advertisement does not meet the standards, the creative model is determined to be invalid, and the control module will issue an alarm. The information data parameters include: click-through rate and positive review rate; The creative model copywriting data parameters include: the matching degree between the copywriting and keywords and the copywriting length.

[0025] In this embodiment, an advertising generation system based on creative features includes a complete system architecture comprising a data acquisition module, an analysis module, an evaluation module, a decision-making module, and a control module. The data acquisition module systematically collects advertising generation data parameters and creative model copy information data parameters from historical periods. The analysis module performs in-depth mining and feature extraction on the collected raw data, identifying strong correlations between data features and the high-efficiency performance of advertising generation. The evaluation module performs rapid and consistent automated evaluation of the results of each batch or each instance of advertising generation based on predetermined standards. The decision-making module further diagnoses whether the creative model meets the standards when advertising generation fails to meet the standards, determining whether it is a random fluctuation in a single generation or a defect in the creative model's capabilities. The control module transforms the results of the analysis and decision-making modules into specific control actions, achieving closed-loop intelligent management of the advertising generation and optimization process.

[0026] Please see Figure 2 The diagram shows a flowchart illustrating the steps involved in generating an advertisement based on creative features according to an embodiment of the present invention. The flowchart includes the following steps: Step S1: Collect information data parameters of ad generation based on creative characteristics of the target platform within the historical period; Step S2: Analyze the feature values ​​of the information data based on the information data parameters; Step S3: Determine whether the generated advertisement meets the standard based on the feature values ​​of the information data; Step S4: If the generated advertisement does not meet the standard, extract the copy information data parameters of the creative model, and analyze the copy information data feature values ​​based on the copy information data parameters. Step S5: Determine whether the creative model meets the standard based on the difference between the feature values ​​of the copywriting information data and the predetermined feature values ​​of the copywriting information data, and propose a processing strategy.

[0027] Specifically, the collection module is used to collect information data parameters of the creative features of the target platform during the historical period, namely click-through rate and positive review rate; the collection module is used to collect copy information data parameters of the creative model, namely the matching degree between copy and keywords and copy length.

[0028] In this embodiment, the click-through rate and positive review rate of the generated advertisements are selected as information data parameters. This allows for the direct capture of the advertisement's attractiveness and user satisfaction from real market feedback, ensuring the business orientation and effectiveness of the evaluation benchmark. For the creative model itself, the matching degree between the copy and keywords and the copy length are selected as copy information data parameters. These provide quantifiable measurement standards for copy quality from two dimensions: semantic relevance and formal standardization. This balances search optimization and user experience. The information data parameters and copy information data parameters together constitute a comprehensive and balanced evaluation basis, enabling subsequent analysis and decision-making to not only gain insights into market performance but also accurately pinpoint optimization directions, thereby driving the entire system to achieve closed-loop, data-driven intelligent evolution.

[0029] Specifically, the feature values ​​of the information data analyzed by the analysis module are determined based on the sum of a first feature-limited representation parameter and a second feature-limited representation parameter, wherein, The first feature defines the characterization parameter as the ratio of the click-through rate to a predetermined click-through rate; The second feature defines the characteristic parameter as the ratio of the positive review rate to the predetermined positive review rate.

[0030] In this embodiment, the predetermined click-through rate and the predetermined positive review rate are both obtained in advance. The benchmark values ​​are derived from the statistical calculation and analysis of the information data parameters generated by the creative characteristics of the target platform within the historical period, so as to provide a stable and unified objective reference standard for advertising quality evaluation.

[0031] In this embodiment, the information data parameters are transformed into comparable information data feature values ​​through normalization processing, which enhances the system's anti-interference ability and stability, and also provides a clear optimization target for the model, laying the foundation for intelligence. The quantifiable framework of Embodiment 1 strongly supports operational automation and scaling, while its flexible architecture makes it easy to expand new evaluation dimensions in the future, thereby achieving a comprehensive improvement in accuracy, robustness and efficiency.

[0032] Specifically, the feature values ​​of the copywriting information data of the creative model analyzed by the analysis module are determined based on the sum of the first copywriting data constraint parameter and the second copywriting data constraint parameter, wherein, The first text data limitation parameter is the ratio of the matching degree between the text and the keywords to the predetermined matching degree between the text and the keywords; The second text data limitation parameter is the ratio of the predetermined text length to the text length.

[0033] In this embodiment, the predetermined matching degree between the text and keywords and the predetermined text length are both obtained in advance. They are benchmark values ​​determined based on high-quality text data of the creative model in the historical period and the experience of business experts, providing a stable and unified objective reference for text quality assessment.

[0034] In this embodiment, the calculation method and benchmark setting of the text information data feature values ​​have significant multiple beneficial effects. By summing the matching ratio between the text and keywords and the inverse ratio of the text length, a comprehensive evaluation system that considers both content relevance and formal standardization is constructed. This overcomes the limitations of single-dimensional evaluation, incentivizing the model to produce highly relevant text closely related to the core theme while effectively constraining the text length to tend towards an ideal range. This reduces the damage to user experience caused by excessive length or brevity, thereby driving the overall optimization of text quality towards precision and conciseness. Furthermore, the predetermined matching degree and length benchmarks are derived from historical high-quality data and expert experience. This embodiment provides a stable and unified objective reference for evaluation, ensuring not only the fairness and comparability of the evaluation results but also enhancing the system's robustness against single parameter fluctuations. This design ultimately achieves a quantifiable, interpretable, and business-oriented evaluation closed loop, providing clear, reliable, and precise guidance for the automated iteration and optimization of creative models.

[0035] Please see Figure 3 As shown, this is a logic diagram for determining whether an advertisement generation conforms to the standard in an embodiment of the present invention. The logic for determining whether an advertisement generation conforms to the standard in an embodiment of the present invention includes: Collect information and data parameters on ad generation based on creative characteristics of the target platform within a historical period, including click-through rate and positive review rate; Analyze the feature values ​​of the information data based on the aforementioned information data parameters; If the information data feature value is greater than the predetermined information data feature threshold, the advertisement generation is determined to meet the standard. If the information data feature value is less than or equal to the predetermined information data feature threshold, the advertisement generation is deemed not to meet the standard.

[0036] Specifically, the evaluation module determines that the condition for the advertisement generation to meet the standard is that the information data feature value is greater than the predetermined information data feature threshold.

[0037] In this embodiment, the predetermined information data adjustment threshold is obtained in advance. The system collects all information data feature values ​​of the creative features of the advertisements generated by the target platform within a historical period, calculates their average value, and determines it as the predetermined information data adjustment threshold. By comparing and judging the predetermined information data feature threshold, an objective and unified anomaly judgment standard is established.

[0038] Specifically, the evaluation module determines that the condition for an advertisement not meeting the standard is that the information data feature value is less than or equal to a predetermined information data feature threshold.

[0039] In this embodiment, the predetermined information data adjustment threshold is obtained in advance. The system collects all information data feature values ​​of the creative features of the target platform's advertisements within a historical period, calculates their average value, and determines it as the predetermined information data feature threshold. By comparing and judging the predetermined information data feature threshold, an objective and unified anomaly judgment standard is established.

[0040] In this embodiment, by collecting the information data feature values ​​of all advertisements within a historical period and calculating the average value as a threshold, an objective and unified evaluation benchmark based on actual business performance is established. This effectively eliminates the arbitrariness of subjective judgment and ensures the fairness and consistency of the screening criteria. By directly comparing the information data feature values ​​of the advertisements to be evaluated with this threshold, the system can quickly and accurately distinguish between advertisements that meet the standards and those that do not, achieving efficient and batch automated initial screening and greatly improving operational efficiency. The design of Embodiment 1 not only ensures the overall quality of online advertisements and maintains the basic effectiveness of the campaign, but more importantly, it provides a clear signal for subsequent processes: compliant advertisements can quickly enter the campaign stage, while non-compliant advertisements trigger in-depth diagnostic and optimization mechanisms. This constructs a refined operational closed loop of evaluation, screening, and triage, laying a solid foundation for the adaptive optimization and efficient resource allocation of the entire system.

[0041] Specifically, the decision module determines that the creative model generated by an advertisement that does not meet the standard meets the standard if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is greater than a predetermined difference threshold.

[0042] In this embodiment, the difference threshold between the text information data feature value and the predetermined text information data feature value is obtained in advance. The system collects the difference between all text information data feature values ​​of the creative model of the target platform within a historical period and the predetermined text information data feature value, calculates their average value, determines the difference threshold between the text information data feature value and the predetermined text information data feature value, and establishes an objective and unified anomaly judgment standard by comparing and judging the predetermined information data feature threshold.

[0043] Specifically, the decision module determines that the creative model generated by the non-compliant advertisement does not meet the standard if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is less than or equal to the predetermined difference threshold.

[0044] In this embodiment, the difference threshold between the text information data feature value and the predetermined text information data feature value is obtained in advance. The system collects the difference between all text information data feature values ​​of the creative model of the target platform within a historical period and the predetermined text information data feature value, calculates their average value, determines the difference threshold between the text information data feature value and the predetermined text information data feature value, and establishes an objective and unified anomaly judgment standard by comparing and judging the predetermined information data feature threshold.

[0045] In this embodiment, a secondary decision-making mechanism based on a predetermined difference threshold provides the advertising generation system with crucial refined screening and resource optimization capabilities. After an ad fails the initial screening, the difference between its text information data feature values ​​and the predetermined text information data feature threshold is further compared. An objective difference threshold derived from historical data is used for judgment, achieving efficient differentiation between potential and inferior models. When the difference in text information data features exceeds the predetermined difference threshold, it indicates that the copywriting quality of the creative model is significantly better than the historical average, possessing high optimization value. The system can then determine that it meets the standards and initiate a special optimization process, effectively rescuing high-quality creatives that failed the initial screening due to other accidental factors and reducing losses from misjudgments. Conversely, it signifies a fundamental defect in the model quality, and the system can promptly terminate subsequent resource investment. This embodiment designs and constructs an intelligent decision-making chain of initial screening, refined judgment, and diversion, which not only significantly improves the accuracy and efficiency of system resource allocation but also ensures, through a data-driven approach, that the creative optimization process always focuses on the most promising direction, thereby strengthening the system's overall adaptability and output efficiency.

[0046] Please see Figure 4 As shown, this is a logic diagram for determining whether a creative model conforms to the standard in an embodiment of the present invention. The process for determining whether a creative model conforms to the standard in an embodiment of the present invention includes: If the generated advertisement does not meet the standard, the copy information data parameters of the creative model are analyzed based on the copy information data parameters to determine the copy information data feature values. If the difference between the feature value of the copywriting information data and the predetermined feature value of the copywriting information data is greater than the predetermined difference threshold, the creative model is determined to meet the standard, and the threshold for adjusting the information data feature is determined. If the difference between the feature value of the copywriting information data and the predetermined feature value of the copywriting information data is less than or equal to the difference threshold, the creative model is determined to be non-compliant with the standard and an alarm is issued.

[0047] Specifically, the condition for the control module to adjust the predetermined information data feature threshold is that the creative model generated by the advertisement that does not meet the standard meets the standard.

[0048] Specifically, the condition for the control module to issue an alarm is that the creative model generated by the non-compliant advertisement does not meet the standards.

[0049] In this embodiment, by introducing predetermined information data feature thresholds for dynamic adjustment and alarm triggering, the system's adaptability, resource utilization efficiency, and risk management level are significantly improved. These measures continuously raise the system's quality threshold, driving the creative model to continuously optimize and surpass historical averages. This effectively reduces model stagnation, ensuring that ad generation performance adapts to market changes and continuously improves, achieving dynamic optimization of the system benchmark. The alarm mechanism triggered when the optimized model still does not meet the standards constitutes a crucial risk control and resource protection barrier. It promptly notifies operations personnel to manually intervene in substandard creatives or directly terminate ineffective optimization processes, preventing continuous waste of computing resources and operating costs. It also effectively intercepts the delivery of low-quality ads, protecting user experience and brand image. These two aspects complement each other: the former drives the system to continuously improve, while the latter ensures timely loss mitigation in operations, together forming a closed-loop intelligent management system that combines progressiveness and robustness.

[0050] The technical solutions of the present invention have been described in conjunction with the embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions resulting from these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An advertising generation system based on creative features, characterized in that, include: The data collection module is used to collect information data parameters of the creative characteristics of the target platform during the historical period, as well as copy information data parameters of the creative model. An analysis module, connected to the acquisition module, is used to analyze the feature values ​​of information data based on information data parameters; Analysis of copywriting information data feature values ​​based on creative model; An evaluation module, which is connected to the analysis module, is used to determine whether the generated advertisement meets the standards based on the feature values ​​of the information data; The decision module, which is connected to the analysis module and the evaluation module, is used to determine whether the creative model generated by the non-standard advertisement meets the standard based on the feature values ​​of the copywriting information data. The control module, which is connected to the decision module, is used to propose corresponding processing strategies based on non-standard advertisement generation: determining to adjust the predetermined information data feature threshold in order to determine to issue an alarm; The information data parameters include: click-through rate and positive review rate; The creative model copywriting data parameters include: the matching degree between the copywriting and keywords and the copywriting length.

2. The advertising generation system based on creative features according to claim 1, characterized in that, The data collection module is used to collect information data parameters of the creative features of the target platform during the historical period, namely click-through rate and positive review rate; the data collection module is used to collect copy information data parameters of the creative model, namely the matching degree between copy and keywords and copy length.

3. The advertising generation system based on creative features according to claim 2, characterized in that, The information data feature values ​​analyzed by the analysis module are determined based on the sum of the first feature-limited representation parameter and the second feature-limited representation parameter, wherein, The first feature defines the characterization parameter as the ratio of the click-through rate generated by the advertisement to the predetermined click-through rate; The second feature defines the characteristic parameter as the ratio of the positive review rate generated by the advertisement to the predetermined positive review rate.

4. The advertising generation system based on creative features according to claim 3, characterized in that, The feature values ​​of the copywriting information data of the creative model analyzed by the analysis module are determined based on the sum of the first copywriting data constraint parameter and the second copywriting data constraint parameter, wherein, The first text data limitation parameter is the ratio of the matching degree between the text and the keywords to the predetermined matching degree between the text and the keywords; The second text data limitation parameter is the ratio of the predetermined text length to the text length.

5. The advertising generation system based on creative features according to claim 4, characterized in that, The evaluation module is used to determine whether an advertisement meets the standard if the information data feature value is greater than a predetermined information data feature threshold.

6. The advertising generation system based on creative features according to claim 5, characterized in that, The evaluation module determines whether an advertisement does not meet the standard if the information data feature value is less than or equal to a predetermined information data feature threshold.

7. The advertising generation system based on creative features according to claim 6, characterized in that, The The decision module determines whether an ad-generated creative model that does not meet the standard meets the standard if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is greater than the predetermined difference threshold.

8. The advertising generation system based on creative features according to claim 7, characterized in that, The The decision module determines whether an ad-generated creative model that does not meet the standard is non-compliant if the difference between the feature value of the copy information data and the predetermined feature value of the copy information data is less than or equal to the predetermined difference threshold.

9. The advertising generation system based on creative features according to claim 8, characterized in that, The The control module adjusts the predetermined information data feature thresholds based on the condition that the creative model generated by the advertisement does not meet the standard.

10. The advertising generation system based on creative features according to claim 9, characterized in that, The control module issues an alarm when the creative model generated by the non-compliant advertisement does not meet the standards.

Images