Big data advertisement putting effect analysis system and method based on cloud platform

Abstract

The invention discloses a big data advertisement putting effect analysis system and method based on a cloud platform, and relates to the technical field of big data analysis. According to advertisement content, advertisement labels are extracted, and advertisement target groups are determined; collecting advertisement putting data and user behavior data of each platform, constructing a data analysis module, and analyzing audience labels of each platform and advertisement conversion data of each label in each time period; constructing an effect prediction model, and predicting the delivery effects of different advertisement labels in different platforms in different time periods; in combination with the indexes, a comprehensive scoring system is established, and a personalized putting suggestion report for the advertisement is generated; the advertisement putting effect is monitored in real time, effect verification and strategy adjustment are carried out through an A / B test, and the optimal advertisement putting effect is ensured. According to the method, the label extraction and the target group analysis are refined, various indexes and data are integrated, the advertisement putting effect is visually displayed, and an optimization strategy is provided.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a big data advertising delivery effect analysis system and method based on a cloud platform. Background Art

[0002] With the popularization of the Internet and mobile Internet, digital advertising has become an important means of corporate marketing. Traditional advertising methods often rely on experience and simple statistical analysis, which makes it difficult to achieve accurate delivery and effect optimization. In recent years, advertisers have been able to obtain a large amount of user behavior data and advertising delivery data, which has made it possible to more refined advertising delivery.

[0003] However, existing advertising effectiveness analysis systems still have the following problems: data from different platforms are often stored in a decentralized manner, lacking a unified data processing and analysis framework; many systems only analyze from one or a few dimensions, ignoring the complex relationship between advertising content, audience characteristics and time factors; most systems use static analysis methods and are unable to monitor and adjust advertising strategies in real time, resulting in delayed advertising effects; existing systems usually provide general analysis results and lack personalized delivery recommendations for specific advertisements and target groups. Summary of the invention

[0004] The purpose of the present invention is to provide a big data advertising effect analysis system and method based on a cloud platform to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: a method for analyzing the effect of advertising delivery based on big data on a cloud platform, the method comprising the steps of:

[0006] S1. Extract advertising tags based on advertising content and identify the target group of the advertisement;

[0007] S2. Collect advertising data and user behavior data from each platform, build a data analysis module, and analyze the audience labels of each platform and the advertising conversion data of each label in each time period;

[0008] S3. Build an effect prediction model to predict the effects of different advertising tags on different platforms at different time periods;

[0009] S4. Combine the indicators to establish a comprehensive scoring system and generate a personalized advertising recommendation report;

[0010] S5. Monitor the effectiveness of advertising in real time, verify the effectiveness and adjust the strategy through A / B testing to ensure the best effect of advertising.

[0011] According to the above scheme, the advertising content includes advertising text, advertising pictures and advertising videos; the advertising tag extracts advertising keywords based on the core content of the advertisement to form a tag, and the core content includes product category, brand name, functional features and product price;

[0012] Based on the ad tag information, analyze the age group, consumption level, region and behavioral characteristics of the ad target group to identify the ad target group.

[0013] According to the above scheme, the advertising data includes display volume, click volume, interaction volume, advertising expenditure, click-through rate and conversion rate; the user behavior data includes the user's behavior after clicking the advertisement, user browsing history, user purchase history, user consumption level, user age group and user region;

[0014] The collected advertising data and user behavior data are formatted uniformly to ensure data consistency; duplicate and invalid data are removed, missing values ​​are filled, and data integrity and accuracy are ensured;

[0015] The audience tags include age group, consumption level, region, and behavioral characteristics;

[0016] The advertising conversion data of each label in each time period, and analyzing the conversion rate of the advertisements in different time periods according to the delivery time of the advertisements of each label;

[0017] The conversion rate is calculated based on the display data, click data and conversion data of the advertisement, and the conversion effect of different advertisement tags; the conversion effect includes the click, registration, browsing and purchase actions taken by the audience after the advertisement is displayed;

[0018] The data analysis model analyzes the conversion effects of different advertising labels on different audience groups in different time periods.

[0019] According to the above solution, the collected advertising data and user behavior data from each platform, as well as the analyzed audience labels of each platform and the advertising conversion data of each label in each time period, are stored in a structured manner on the cloud platform, and the data is efficiently partitioned and indexed to improve subsequent query and analysis performance.

[0020] According to the above scheme, the relationship between the advertising label, the audience group label and the time period is analyzed by using the cross analysis technology;

[0021] Cross-analysis includes cross-analysis between ad tags and audience tags, cross-analysis between ad tags and time periods, and cross-analysis between audience tags and time periods.

[0022] Through cross-analysis, we can identify which advertising tags, audience tags, and time periods have a strong correlation with conversion effects. The cross-analysis results form a set of input features, which serve as input features for the effect prediction model.

[0023] Encode, normalize, and standardize ad tags, audience tags, and time periods according to the above scheme;

[0024] Using multiple linear regression to predict the advertising effect, the formula is as follows:

[0025] y=β0+β1x1+β2x2+…+β p x p +ε;

[0026] Among them, y is the dependent variable, which represents the conversion rate; β0 is the intercept; β1, β2, …, β p Expressed as regression coefficient, the influence of each variable on the dependent variable; x1, x2, ..., x p denoted as an independent variable, representing different influencing factors, including ad labels, audience group labels, and time periods; ε represents an unpredictable random error;

[0027] Using the least squares method, the optimal regression coefficient is found by minimizing the loss function. The formula is as follows:

[0028]

[0029] Where n is the number of samples; y i Expressed as the conversion rate of the i-th sample; x i1 , x i2 , …, x ip is represented as the independent variable of the i-th sample; β1, β2, …, β p Expressed as regression coefficient;

[0030] By minimizing the loss function, the regression model finds the coefficients (β0, β1, β2, …, β p ).

[0031] According to the above scheme, the determination coefficient is used to compare the difference between the actual value and the predicted value to evaluate the fitting effect of the effect prediction model. The formula is as follows:

[0032]

[0033] Among them, R 2 It is expressed as the coefficient of determination. The closer it is to 1, the better the model can explain the changes in the dependent variable. n is the number of samples. y i Represented as the actual value of the i-th sample; Represented as the predicted value of the i-th sample; It is expressed as the mean value of the dependent variable y;

[0034] To prevent overfitting of the effect prediction model, the regularization method Lasso regression is used, and the formula is as follows:

[0035]

[0036] Among them, λ represents the regularization parameter, which is used to control the complexity of the effect prediction model; p represents the total number of features; j represents the index of the regression coefficient, which is used to identify different independent variables.

[0037] According to the above scheme, the indicators include conversion rate, click-through rate, interaction rate, cost-effectiveness of advertising, user profile matching and advertising exposure;

[0038] Each indicator is standardized to eliminate the impression caused by different dimensions. The formula is as follows:

[0039]

[0040] Among them, standardized(z m ) is expressed as the standardized index value; z m Expressed as the original value of the index m; μ m Expressed as the mean of the index m; σ m It is expressed as the standard deviation of the indicator m;

[0041] Each standardized indicator will be assigned a different weight according to its importance, and finally the comprehensive score will be calculated. The formula is as follows:

[0042]

[0043] Among them, score represents the comprehensive score; wm represents the weight of indicator m; k represents the total number of indicators.

[0044] Establish an automated model update mechanism and regularly use new data to retrain the effect prediction model to maintain the prediction accuracy of the effect prediction model.

[0045] According to the above scheme, based on the indicators and comprehensive scores, problem diagnosis is performed and optimization strategies are proposed. The problem diagnosis includes inefficient advertising creativity or materials, mismatched delivery time periods, targeting accuracy problems and unreasonable budget allocation; the optimization strategies include modifying advertising content, adjusting target audiences and optimizing delivery time;

[0046] The delivery recommendation report includes advertising tags, audience tags, advertising delivery effect predictions, indicator values, comprehensive scores, problem diagnosis and optimization strategies.

[0047] A big data advertising effect analysis system based on a cloud platform, the system includes: a data collection module, a data storage module, a data analysis module, a model prediction module and an analysis and display module;

[0048] The data collection module provides a standard API interface to facilitate integration with various platforms; automatically collects advertising data and user behavior data of various platforms; the advertising data includes advertising display volume, click volume, interaction volume, advertising expenditure, click-through rate and conversion rate; the user behavior data includes user click behavior, browsing history, purchase history, consumption level, age group and region; the collected data is transmitted to the data analysis module and the data storage module;

[0049] The data storage module receives the data collected by the data collection module and stores it in a structured manner in the cloud platform; performs efficient partitioning and indexing of the data to improve query performance and support storage and processing of large-scale data;

[0050] The data analysis module receives the data collected by the data collection module, performs pre-processing, and then performs in-depth analysis on the collected advertising data and user behavior data, analyzes the audience labels of the advertisements and the advertising conversion effects of the advertisement labels in different time periods, and finds out the relationship between advertising and audience characteristics;

[0051] The model prediction module constructs an effect prediction model, uses a multivariate linear regression model, cross-analysis results, and standardized data to predict advertising effects, optimizes the effect prediction model through the least squares method and regularization method, and outputs a comprehensive score;

[0052] The analysis and display module generates a personalized report on advertising delivery based on the data analysis module and the model prediction module, and displays it to the user in a visual manner; the personalized report includes advertising tags, audience tags, advertising delivery effect predictions, indicator values, comprehensive scores, problem diagnosis and optimization strategies; and provides an interactive report interface to allow users to customize the viewing of data and analysis results in different dimensions.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. The refined label extraction and target group analysis of the present invention make the analysis of advertising effect more accurate and specific;

[0055] 2. The present invention uses cross analysis technology and multiple linear regression model to improve the accuracy and generalization ability of advertising effect prediction

[0056] 3. The present invention integrates various indicators and data to complete comprehensive scoring, intuitively displays the effect of advertising, and provides optimization strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a flow chart of the steps of the big data advertising effect analysis method based on the cloud platform of the present invention;

[0058] Figure 2 It is a structural schematic diagram of the big data advertising delivery effect analysis system based on the cloud platform of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0060] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a method for analyzing the effect of advertising delivery based on big data on a cloud platform;

[0061] Extract advertising tags based on advertising content and identify the target group of the advertisement;

[0062] Ad content: ad copy, images and video for a new summer clothing line;

[0063] Advertising tags: Product category: summer clothing; Brand name: XYZ; Features: light and breathable, fashionable design; Product price: medium price;

[0064] Target group of advertisement: Age group: 18-35 years old; Consumption level: upper-middle level; Region: first-tier cities; Behavioral characteristics: like fashion trends and often shop online.

[0065] Collect advertising data and user behavior data from each platform. The advertising data includes impressions, clicks, interactions, advertising expenditures, click-through rates, and conversion rates. The user behavior data includes user behavior after clicking on an ad, user browsing history, user purchase history, user consumption level, user age group, and user region.

[0066] Advertising data of a certain platform: impressions: 100,000 times; clicks: 2,000 times; interactions: 500 times; advertising expenditure: 10,000 yuan; click rate: 2%; conversion rate: 0.5%;

[0067] User behavior data: user behavior after clicking: browsing detail page, adding to shopping cart, purchasing; user browsing history: frequently browsing fashion products; user purchase history: purchased similar products in the past month; user consumption level: the average purchase amount is 500 yuan each time; user age group: mainly concentrated in 20-30 years old; user region: Beijing, Shanghai, Guangzhou;

[0068] Advertising conversion data of each label in each time period: Time period 1 (9:00-12:00): 10,000 impressions, 200 clicks, and 10 conversions; Time period 2 (12:00-15:00): 15,000 impressions, 300 clicks, and 15 conversions; Time period 3 (15:00-18:00): 20,000 impressions, 400 clicks, and 20 conversions; Time period 4 (18:00-21:00): 25,000 impressions, 500 clicks, and 25 conversions; Time period 5 (21:00-24:00): 30,000 impressions, 600 clicks, and 30 conversions.

[0069] The collected advertising data and user behavior data from each platform, as well as the analyzed audience tags of each platform and the advertising conversion data of each tag in each time period, are stored in a structured manner on the cloud platform, and the data is efficiently partitioned and indexed to improve subsequent query and analysis performance.

[0070] Input features: Advertisement tags: summer clothing, XYZ, light and breathable, medium price; Audience tags: 20-30 years old, upper middle class, first-tier cities, like fashion trends; Time periods: 9:00-12:00, 12:00-15:00, 15:00-18:00, 18:00-21:00, 21:00-24:00;

[0071] Using multiple linear regression to predict the advertising effect, the formula is as follows:

[0072] y=β0+β1x1+β2x2+…+β p x p +ε;

[0073] Among them, y is the dependent variable, which represents the conversion rate; β0 is the intercept; β1, β2, …, β p Expressed as regression coefficient, the influence of each variable on the dependent variable; x1, x2, ..., x p denoted as an independent variable, representing different influencing factors, including ad labels, audience group labels, and time periods; ε represents an unpredictable random error;

[0074] Using the least squares method, the optimal regression coefficient is found by minimizing the loss function. The formula is as follows:

[0075]

[0076] Where n is the number of samples; y i Expressed as the conversion rate of the i-th sample; x i1 , x i2 , …, x ip is represented as the independent variable of the i-th sample; β1, β2, …, β p Expressed as regression coefficient;

[0077] By minimizing the loss function, the regression model finds the coefficients (β0, β1, β2, …, β p ).

[0078] According to the above scheme, the determination coefficient is used to compare the difference between the actual value and the predicted value to evaluate the fitting effect of the effect prediction model. The formula is as follows:

[0079]

[0080] Among them, R 2 It is expressed as the coefficient of determination. The closer it is to 1, the better the model can explain the changes in the dependent variable. n is the number of samples. y i Represented as the actual value of the i-th sample; Represented as the predicted value of the i-th sample; It is expressed as the mean value of the dependent variable y;

[0081] To prevent overfitting of the effect prediction model, the regularization method Lasso regression is used, and the formula is as follows:

[0082]

[0083] Among them, λ represents the regularization parameter, which is used to control the complexity of the effect prediction model; p represents the total number of features; j represents the index of the regression coefficient, which is used to identify different independent variables.

[0084] According to the above scheme, the indicators include conversion rate, click-through rate, interaction rate, cost-effectiveness of advertising, user profile matching and advertising exposure;

[0085] Conversion rate: 0.5%; Click-through rate: 2%; Interaction rate: 0.5%; Advertising cost-effectiveness: 10,000 yuan / 50 conversions = 200 yuan / time; User portrait matching: high; Advertising exposure: 100,000 times;

[0086] Each indicator is standardized to eliminate the impression caused by different dimensions. The formula is as follows:

[0087]

[0088] Among them, standardized(z m ) is expressed as the standardized index value; z m Expressed as the original value of the index m; μ m Expressed as the mean of the index m; σ m It is expressed as the standard deviation of the indicator m;

[0089] Each standardized indicator will be assigned a different weight according to its importance, and finally the comprehensive score will be calculated. The formula is as follows:

[0090]

[0091] Among them, score represents the comprehensive score; wm represents the weight of indicator m; k represents the total number of indicators

[0092] Based on the indicators and comprehensive scores, problem diagnosis is performed and optimization strategies are proposed. The problem diagnosis includes inefficient advertising creativity or materials, mismatched delivery time periods, targeting accuracy issues, and unreasonable budget allocation; the optimization strategies include modifying advertising content, adjusting target audiences, and optimizing delivery time;

[0093] The delivery recommendation report includes ad tags, audience tags, ad delivery effect prediction, indicator values, comprehensive scores, problem diagnosis and optimization strategies;

[0094] Optimization strategy: modify the content of the advertisement: add limited-time discount information to increase appeal; adjust the target audience: refine the target audience and increase targeted delivery to specific interest groups; optimize the delivery time: concentrate resources on the time period of 18:00-21:00 for delivery; budget reallocation: reduce the budget on platforms with poor performance and increase the budget on platforms with high conversion rates;

[0095] Monitor advertising effectiveness in real time, verify effectiveness and adjust strategies through A / B testing to ensure optimal advertising results.

[0096] The present invention provides another technical solution, a big data advertising effect analysis system based on a cloud platform, the system comprising: a data acquisition module, a data storage module, a data analysis module, a model prediction module and an analysis and display module;

[0097] The data collection module provides a standard API interface to facilitate integration with various platforms; automatically collects advertising data and user behavior data of various platforms; the advertising data includes advertising display volume, click volume, interaction volume, advertising expenditure, click-through rate and conversion rate; the user behavior data includes user click behavior, browsing history, purchase history, consumption level, age group and region; the collected data is transmitted to the data analysis module and the data storage module;

[0098] The data storage module receives the data collected by the data collection module and stores it in a structured manner in the cloud platform; performs efficient partitioning and indexing of the data to improve query performance and support storage and processing of large-scale data;

[0099] The data analysis module receives the data collected by the data collection module, performs pre-processing, and then performs in-depth analysis on the collected advertising data and user behavior data, analyzes the audience labels of the advertisements and the advertising conversion effects of the advertisement labels in different time periods, and finds out the relationship between advertising and audience characteristics;

[0100] The model prediction module constructs an effect prediction model, uses a multivariate linear regression model, cross-analysis results, and standardized data to predict advertising effects, optimizes the effect prediction model through the least squares method and regularization method, and outputs a comprehensive score;

[0101] The analysis and display module generates a personalized report on advertising delivery based on the data analysis module and the model prediction module, and displays it to the user in a visual manner; the personalized report includes advertising tags, audience tags, advertising delivery effect predictions, indicator values, comprehensive scores, problem diagnosis and optimization strategies; and provides an interactive report interface to allow users to customize the viewing of data and analysis results in different dimensions.

[0102] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A big data advertising effect analysis method based on a cloud platform, characterized by: The method comprises the steps of: S1. Extract advertising tags based on advertising content and identify the target group of the advertisement; S2. Collect advertising data and user behavior data from each platform, build a data analysis module, and analyze the audience labels of each platform and the advertising conversion data of each label in each time period; S3. Build an effect prediction model to predict the effects of different advertising tags on different platforms at different time periods; S4. Combine the indicators to establish a comprehensive scoring system and generate a personalized advertising recommendation report; S5. Monitor the effectiveness of advertising in real time, verify the effectiveness and adjust the strategy through A / B testing to ensure the best effect of advertising.

2. The big data advertising delivery effect analysis method based on the cloud platform according to claim 1 is characterized by: The advertisement content includes advertisement text, advertisement pictures and advertisement videos; the advertisement tag extracts advertisement keywords according to the core content of the advertisement to form a tag, and the core content includes product category, brand name, functional features and product price; Based on the ad tag information, analyze the age group, consumption level, region and behavioral characteristics of the ad target group to identify the ad target group.

3. The big data advertising effect analysis method based on the cloud platform according to claim 1 is characterized by: The advertising data includes the number of impressions, clicks, interactions, advertising expenditures, click-through rates and conversion rates; the user behavior data includes the user's behavior after clicking on the advertisement, user browsing history, user purchase history, user consumption level, user age group and user region; Format the collected advertising data and user behavior data in a unified manner to ensure data consistency; Remove duplicate and invalid data, fill in missing values, and ensure data integrity and accuracy; The audience tags include age group, consumption level, region, and behavioral characteristics; The advertising conversion data of each label in each time period, and analyzing the conversion rate of the advertisements in different time periods according to the delivery time of the advertisements of each label; The conversion rate is calculated based on the display data, click data and conversion data of the advertisement, and the conversion effect of different advertisement tags; the conversion effect includes the click, registration, browsing and purchase actions taken by the audience after the advertisement is displayed; The data analysis model analyzes the conversion effects of different advertising labels on different audience groups in different time periods.

4. The big data advertising effect analysis method based on the cloud platform according to claim 3 is characterized by: The collected advertising data and user behavior data from each platform, as well as the analyzed audience tags of each platform and the advertising conversion data of each tag in each time period, are stored in a structured manner on the cloud platform, and the data is efficiently partitioned and indexed to improve subsequent query and analysis performance.

5. The big data advertising effect analysis method based on the cloud platform according to claim 1 is characterized by: Use cross-analysis technology to analyze the relationship between advertising tags, audience tags, and time periods; Cross-analysis includes cross-analysis between ad tags and audience tags, cross-analysis between ad tags and time periods, and cross-analysis between audience tags and time periods. Through cross-analysis, we can identify which ad tags, audience tags, and time periods have a strong correlation with conversion effects. The cross-analysis results form a set of input features. As input features of the effect prediction model.

6. The big data advertising effect analysis method based on the cloud platform according to claim 1 is characterized by: Encode, normalize, and standardize ad tags, audience tags, and time periods; Using multiple linear regression to predict the advertising effect, the formula is as follows: y=β0+β1x1+β2x2+…+β p x p +e; Among them, y is the dependent variable, which represents the conversion rate; β0 is the intercept; β1, β2, …, β p Expressed as regression coefficient, the influence of each variable on the dependent variable; x1, x2, ..., x p denoted as an independent variable, representing different influencing factors, including ad labels, audience group labels, and time periods; ε represents an unpredictable random error; Using the least squares method, the optimal regression coefficient is found by minimizing the loss function. The formula is as follows: Where n is the number of samples; y i Expressed as the conversion rate of the i-th sample; x i1 , x i2 , …, x ip is represented as the independent variable of the i-th sample; β1, β2, …, β p Expressed as regression coefficient; By minimizing the loss function, the regression model finds the coefficients (β0, β1, β2, …, β p ).

7. The big data advertising delivery effect analysis method based on the cloud platform according to claim 5 is characterized by: The determination coefficient is used to compare the difference between the actual value and the predicted value and evaluate the fitting effect of the effect prediction model. The formula is as follows: Among them, R 2 is represented by the coefficient of determination; n is the sample size; y i Represented as the actual value of the i-th sample; Represented as the predicted value of the i-th sample; It is expressed as the mean value of the dependent variable y; To prevent overfitting of the effect prediction model, the regularization method Lasso regression is used, and the formula is as follows: Among them, λ represents the regularization parameter, which is used to control the complexity of the effect prediction model; p represents the total number of features; j represents the index of the regression coefficient, which is used to identify different independent variables.

8. The big data advertising delivery effect analysis method based on the cloud platform according to claim 1 is characterized by: The indicators include conversion rate, click-through rate, interaction rate, advertising cost-effectiveness, user profile matching and advertising exposure; Each indicator is standardized to eliminate the impression caused by different dimensions. The formula is as follows: Among them, standardized(z m ) is expressed as the standardized index value; z m Represented as the original value of the index m; μ m Expressed as the mean of the index m; σ m It is expressed as the standard deviation of the indicator m; Each standardized indicator will be assigned a different weight according to its importance, and finally the comprehensive score will be calculated. The formula is as follows: Among them, score represents the comprehensive score; wm represents the weight of indicator m; k represents the total number of indicators.

9. The big data advertising effect analysis method based on the cloud platform according to claim 1 is characterized by: Based on the indicators and comprehensive scores, problem diagnosis is performed and optimization strategies are proposed. The problem diagnosis includes inefficient advertising creativity or materials, mismatched delivery time periods, targeting accuracy issues, and unreasonable budget allocation; the optimization strategies include modifying advertising content, adjusting target audiences, and optimizing delivery time; The delivery recommendation report includes advertising tags, audience tags, advertising delivery effect predictions, indicator values, comprehensive scores, problem diagnosis and optimization strategies.

10. A big data advertising effect analysis system based on a cloud platform, characterized by: The system includes: a data acquisition module, a data storage module, a data analysis module, a model prediction module and an analysis and display module; The data collection module integrates the APIs of various platforms to automatically collect advertising data and user behavior data of various platforms; the advertising data includes advertising display volume, click volume, interaction volume, advertising expenditure, click-through rate and conversion rate; the user behavior data includes the user's click behavior, browsing history, purchase history, consumption level, age group and region; the collected data is transmitted to the data analysis module and the data storage module; The data storage module receives the data collected by the data collection module and stores it in a structured manner in the cloud platform; partitions and indexes the data efficiently to improve query performance and support storage and processing of large-scale data; The data analysis module receives the data collected by the data collection module, performs pre-processing, and then performs in-depth analysis on the collected advertising data and user behavior data, analyzes the audience labels of the advertisements and the advertising conversion effects of the advertisement labels in different time periods, and finds out the relationship between advertising and audience characteristics; The model prediction module constructs an effect prediction model, uses a multivariate linear regression model, cross-analysis results, and standardized data to predict advertising effects, optimizes the effect prediction model through the least squares method and regularization method, and outputs a comprehensive score; The analysis and display module generates a personalized report on advertising delivery based on the data analysis module and the model prediction module, and displays it to the user in a visual manner; the personalized report includes advertising tags, audience tags, advertising delivery effect predictions, indicator values, comprehensive scores, problem diagnosis and optimization strategies.

Images