New media advertisement putting and monitoring integrated system

Abstract

The invention discloses a new media advertisement putting and monitoring integrated system, and the system comprises a user portrait construction module which obtains the browsing behavior, geographic position and social interaction data of a user at a mobile terminal, a PC terminal and an intelligent terminal through a multi-source data collection unit, and generates a dynamically updated user interest tag; the intelligent advertisement putting engine is used for matching advertisement contents with user portraits in real time based on a deep reinforcement learning algorithm and supporting programmed cross-platform bidding and private market direct putting; according to the invention, through multi-dimensional effect evaluation, emotion analysis and user life cycle value prediction are introduced, and a comprehensive evaluation system is constructed; data security is enhanced, and through homomorphic encryption and block chain technologies, user privacy is guaranteed and a non-tampering effect auditing voucher is provided. The customer obtaining cost of advertisers can be obviously reduced, and the rate of return on investment is improved.

Description

Technical Field

[0001] This invention belongs to the field of advertising placement technology, specifically relating to an integrated system for new media advertising placement and monitoring. Background Technology

[0002] Traditional advertising placement and monitoring systems suffer from numerous shortcomings: ad placement and performance monitoring are conducted on separate platforms, resulting in significant data synchronization delays and delayed optimization decisions, failing to meet the real-time response demands of the new media environment; the lack of effective abnormal traffic identification mechanisms leads to wasted advertisers' budgets due to fraudulent activities such as fake clicks and bot traffic; strategy optimization relies on human experience, especially when conducting cross-channel campaigns, making it difficult to quickly respond to changes in user behavior and fluctuations in the competitive environment, resulting in low campaign efficiency; user behavior data is scattered across multiple platforms, limiting traditional tracking technologies and causing low accuracy in cross-device user identification, leading to biased attribution analysis; and evaluation metrics are limited, lacking quantitative assessments of soft indicators such as brand awareness and user sentiment, making it difficult to comprehensively measure advertising value. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to overcome the existing defects and provide an integrated system for new media advertising placement and monitoring to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an integrated system for new media advertising placement and monitoring, comprising: User profile building module: Acquires user browsing behavior, geographical location, and social interaction data on mobile devices, PCs, and smart terminals through multi-source data collection units, and generates dynamically updated user interest tags; Intelligent Ad Delivery Engine: Based on deep reinforcement learning algorithms, it matches ad content with user profiles in real time, and supports programmatic cross-platform bidding and direct delivery to private marketplaces; End-to-end monitoring module: Integrates SDK tracking, API interfaces and no-code tracking technology to collect real-time data on ad impressions, click-through rates, and conversion paths, and simultaneously tracks user behavior across devices; Data analysis and performance evaluation module: Construct a multi-dimensional evaluation model that includes brand awareness, conversion funnel, and user retention rate, and use the Shapley value algorithm for cross-channel attribution analysis; Closed-loop optimization control module: Dynamically adjusts the delivery strategy based on real-time monitoring data to achieve automated iterative optimization of advertising budget allocation, creative content, and delivery time periods.

[0005] Preferably, the user profile building module further includes: Cross-platform data cleaning unit: performs noise reduction, deduplication, and standardization processing on heterogeneous data from mobile applications, web logs, and e-commerce transaction systems; Real-time interest prediction submodule: Utilizes LSTM neural network to analyze user behavior sequences, dynamically updates interest tag weights, and generates an interest decay model for the next 7 days.

[0006] Preferably, the cross-platform data cleaning unit is configured with: A knowledge graph-based data association engine identifies the anonymous IDs of the same user on different platforms and establishes mapping relationships. The data quality verification rule base automatically repairs or removes outliers, missing values, and conflicting data.

[0007] Preferably, the intelligent advertising delivery engine includes: Dynamic Creative Generator: Automatically synthesizes advertising creatives adapted to different terminal resolutions and scenarios based on generative adversarial networks; Bidding strategy optimizer: Employs a multi-armed slot machine algorithm to adjust bidding parameters based on historical bidding success rates and real-time competition intensity.

[0008] Preferably, the end-to-end monitoring module further includes: Abnormal Traffic Detection Unit: Uses Gaussian mixture model to perform cluster analysis on clickstream data to identify bot traffic, proxy IPs, and click farm behavior; Cross-device attribution unit: Combining device fingerprinting technology, Wi-Fi probes, and cookie mapping, it associates the advertising interaction sequences of the same user across mobile phones, tablets, and smart TVs.

[0009] Preferably, the abnormal traffic detection unit is configured with: A click fraud detection model based on time series analysis can detect short-term high-frequency clicks and non-human operation trajectories, such as uniform-speed swiping. Traffic blacklist / whitelist database, which is synchronized in real time with malicious IPs and device IDs marked by third-party anti-fraud platforms.

[0010] Preferably, the data analysis and effect evaluation module includes: User lifetime value prediction model: The Cox proportional hazards regression algorithm is used to predict the long-term value of users by combining user activity and consumption frequency. Sentiment Analysis Unit: Analyzes user comments using the BERT natural language processing model to quantify the impact of advertising on brand sentiment.

[0011] Preferably, the closed-loop optimization control module performs the following functions: Budget dynamic allocator: Based on a linear programming algorithm, it allocates budget weights according to the real-time ROI of each channel; Intelligent alarm system: When the click-through rate or conversion rate deviates from the predicted value by more than 15%, it triggers an emergency adjustment of the advertising strategy; A / B testing framework: Automatically generates combinations of ad creative variations, such as copy, color scheme, and CTA buttons, and selects the optimal solution after parallel testing.

[0012] Preferred options also include: Blockchain Evidence Storage Unit: Stores ad exposure, click and conversion data on the blockchain to generate an immutable performance verification certificate; Data visualization dashboard: Provides heatmaps, conversion funnels, and Sankey diagrams of user paths for real-time campaign data, supporting multi-dimensional drill-down analysis.

[0013] Preferably, the system ensures data security through the following methods: Homomorphic encryption technology is used to encrypt user privacy data, such as device ID and geolocation. Set up a tiered data access permission mechanism to restrict advertisers to only query aggregated statistical results and not to obtain original user information.

[0014] Compared with existing technologies, the present invention provides an integrated system for new media advertising placement and monitoring, which has the following beneficial effects: This invention achieves rapid synchronization of delivery and monitoring data through real-time closed-loop optimization, significantly shortening the response time for strategy adjustments and improving the timeliness of dynamic bidding and creative optimization. It also features precise anomaly interception, employing advanced models and technologies to effectively identify abnormal traffic and reduce advertisers' ineffective spending; intelligent strategy iteration, based on deep learning and generative adversarial networks, automatically optimizes ad conversion rates and selects the optimal combination of creative materials; cross-platform data fusion utilizes multiple technologies to improve the accuracy of cross-device user identification and scientifically quantify the contribution of multiple touchpoints; multi-dimensional performance evaluation introduces sentiment analysis and user lifetime value prediction to construct a comprehensive evaluation system; and enhanced data security uses homomorphic encryption and blockchain technology to protect user privacy and provide tamper-proof performance audit credentials. This significantly reduces advertisers' customer acquisition costs and improves ROI. Detailed Implementation

[0015] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] This invention provides a technical solution: an integrated system for new media advertising placement and monitoring, comprising: User profile building module: Acquires user browsing behavior, geographical location, and social interaction data on mobile devices, PCs, and smart terminals through multi-source data collection units, and generates dynamically updated user interest tags; Intelligent Ad Delivery Engine: Based on deep reinforcement learning algorithms, it matches ad content with user profiles in real time, and supports programmatic cross-platform bidding and direct delivery to private marketplaces; End-to-end monitoring module: Integrates SDK tracking, API interfaces and no-code tracking technology to collect real-time data on ad impressions, click-through rates, and conversion paths, and simultaneously tracks user behavior across devices; Data analysis and performance evaluation module: Construct a multi-dimensional evaluation model that includes brand awareness, conversion funnel, and user retention rate, and use the Shapley value algorithm for cross-channel attribution analysis; Closed-loop optimization control module: Dynamically adjusts the delivery strategy based on real-time monitoring data to achieve automated iterative optimization of advertising budget allocation, creative content, and delivery time periods.

[0017] In this invention, preferably, the user profile construction module further includes: Cross-platform data cleaning unit: performs noise reduction, deduplication, and standardization processing on heterogeneous data from mobile applications, web logs, and e-commerce transaction systems; Real-time interest prediction submodule: Utilizes LSTM neural network to analyze user behavior sequences, dynamically updates interest tag weights, and generates an interest decay model for the next 7 days.

[0018] In this invention, preferably, the cross-platform data cleaning unit is configured with: A knowledge graph-based data association engine identifies the anonymous IDs of the same user on different platforms and establishes mapping relationships. The data quality verification rule base automatically repairs or removes outliers, missing values, and conflicting data.

[0019] In this invention, preferably, the intelligent advertising delivery engine includes: Dynamic Creative Generator: Automatically synthesizes advertising creatives adapted to different terminal resolutions and scenarios based on generative adversarial networks; Bidding strategy optimizer: Employs a multi-armed slot machine algorithm to adjust bidding parameters based on historical bidding success rates and real-time competition intensity.

[0020] In this invention, preferably, the end-to-end monitoring module further includes: Abnormal Traffic Detection Unit: Uses Gaussian mixture model to perform cluster analysis on clickstream data to identify bot traffic, proxy IPs, and click farm behavior; Cross-device attribution unit: Combining device fingerprinting technology, Wi-Fi probes, and cookie mapping, it associates the advertising interaction sequences of the same user across mobile phones, tablets, and smart TVs.

[0021] In this invention, preferably, the abnormal flow detection unit is configured with: A click fraud detection model based on time series analysis can detect short-term high-frequency clicks and non-human operation trajectories, such as uniform-speed swiping. Traffic blacklist / whitelist database, which is synchronized in real time with malicious IPs and device IDs marked by third-party anti-fraud platforms.

[0022] In this invention, preferably, the data analysis and effect evaluation module includes: User lifetime value prediction model: The Cox proportional hazards regression algorithm is used to predict the long-term value of users by combining user activity and consumption frequency. Sentiment Analysis Unit: Analyzes user comments using the BERT natural language processing model to quantify the impact of advertising on brand sentiment.

[0023] In this invention, preferably, the closed-loop optimization control module performs the following functions: Budget dynamic allocator: Based on a linear programming algorithm, it allocates budget weights according to the real-time ROI of each channel; Intelligent alarm system: When the click-through rate or conversion rate deviates from the predicted value by more than 15%, it triggers an emergency adjustment of the advertising strategy; A / B testing framework: Automatically generates combinations of ad creative variations, such as copy, color scheme, and CTA buttons, and selects the optimal solution after parallel testing.

[0024] Preferably, this invention further includes: Blockchain Evidence Storage Unit: Stores ad exposure, click and conversion data on the blockchain to generate an immutable performance verification certificate; Data visualization dashboard: Provides heatmaps, conversion funnels, and Sankey diagrams of user paths for real-time campaign data, supporting multi-dimensional drill-down analysis.

[0025] In this invention, preferably, the system ensures data security through the following methods: Homomorphic encryption technology is used to encrypt user privacy data, such as device ID and geolocation. Set up a tiered data access permission mechanism to restrict advertisers to only query aggregated statistical results and not to obtain original user information.

[0026] Example 1: A new media advertising placement and monitoring integrated system includes the following steps: Step 1: Data Collection: User A watches basketball videos on Douyin and searches for "basketball shoes" on JD.com. The system associates this cross-platform behavior with the device fingerprint (IMEI + MAC address). The user profile module generates tags: {Sports Preference: Basketball, Consumption Level: High, Recent Shopping Intention: Footwear, Interest Decay Coefficient: 0.8}. Step 2, Intelligent Targeting: The dynamic creative generator calls the GAN network to synthesize vertical videos adapted for Douyin and e-commerce banner images; the bidding strategy optimizer bids at 0.5 yuan / CPM in the RTB market and successfully displays customized ads when user A is browsing Douyin. Step 3, Performance Tracking: The monitoring module records that after user A clicks on the ad, they complete the purchase across devices on the PC; the attribution analysis module calculates the contribution of Douyin exposure using Shapley values, allocating 62% to the ad and 38% to e-commerce search ads. Step 4, Closed-loop optimization: The budget allocator increases the budget weight of the Douyin channel from 30% to 45%; the system automatically generates 3 sets of new materials (highlighting limited editions, discount information, and celebrity endorsements), and selects the option with the highest click-through rate after A / B testing; Technical results: Cross-device attribution accuracy improved to 89%; dynamic bidding strategy reduced CPC costs by 28%; conversion rate increased by 41% through creative optimization.

[0027] Example 2: A new media advertising placement and monitoring integrated system includes the following steps: Step 1, Anomaly Detection: The real-time computing engine detected 150 clicks from the same IP segment within 2 minutes, and all of them were from emulator devices; the traffic analysis unit extracted mouse trajectory data and detected that 98% of the clicks were concentrated within ±5 pixels of the center point of the ad area (non-human operation characteristics). Step 2, Model Determination: Gaussian Mixture Model (GMM) cluster analysis shows that the standard deviation of the click interval for this traffic cluster is only 0.2 seconds (normal users click at 8 seconds, as shown in Step 3); the system determines this to be bot traffic with a confidence level of 99.3%, and updates the blacklist database accordingly. Step 3, Remedial Measures: Immediately suspend advertising to this IP segment and remove abnormal click data from the billing system (a total of 23,458 fraudulent clicks were blocked); the blockchain evidence storage unit generates a verifiable chain of evidence containing timestamps and traffic characteristics; Step 4, Strategy Adjustment: The optimization module automatically reduces the bidding weight of open proxy traffic by 70%; and adds verification rules for device sensor data (gyroscope, GPS). Technical results: The accuracy rate of identifying fake traffic reached 98.7%; it helped advertisers avoid losing 157,000 yuan of their budget; and the response time for abnormal events was shortened to 200 milliseconds.

[0028] Example 3: A new media advertising placement and monitoring integrated system includes the following steps: Step 1: Creative Generation: Based on historical high-converting materials, the GAN network generates 20 creative combinations (comparison of different makeup effects, ingredient analysis, KOL recommendations); the sentiment analysis unit pre-screens materials with a positive sentiment tendency of ≥80% (such as "12-hour makeup lasting" copywriting + laboratory comparison video). Step 2, Launch Testing: The A / B testing framework displays different creative combinations to 50,000 target users within 24 hours; the monitoring module simultaneously collects post-click behavior: page dwell time, comment sentiment value, and add-to-cart rate; Step 3, Results Evaluation: The data analysis module identifies the optimal solution: ingredient analysis materials increased the add-to-cart rate by 27%, but KOL-recommended materials had a higher user retention rate; the LTV prediction model suggests prioritizing KOL content for users with high spending potential. Step 4: Strategy Execution For users whose "ingredient-focused" tag weight is >0.6, push lab comparison materials; for users with social interaction frequency >3 times / day, show KOL tutorial videos; the system automatically updates the creative library weekly and eliminates materials with CTR below average; Technical results: The percentage of positive sentiment in user reviews increased from 68% to 89%; the 30-day repurchase rate increased by 33%; and the efficiency of creative content generation was 20 times higher than that of manual design.

[0029] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A new media advertising placement and monitoring integrated system, characterized in that, include: User profile building module: Acquires user browsing behavior, geographical location, and social interaction data on mobile devices, PCs, and smart terminals through multi-source data collection units, and generates dynamically updated user interest tags; Intelligent Ad Delivery Engine: Based on deep reinforcement learning algorithms, it matches ad content with user profiles in real time, and supports programmatic cross-platform bidding and direct delivery to private marketplaces; End-to-end monitoring module: Integrates SDK tracking, API interfaces and no-code tracking technology to collect real-time data on ad impressions, click-through rates, and conversion paths, and simultaneously tracks user behavior across devices; Data analysis and performance evaluation module: Construct a multi-dimensional evaluation model that includes brand awareness, conversion funnel, and user retention rate, and use the Shapley value algorithm for cross-channel attribution analysis; Closed-loop optimization control module: Dynamically adjusts the delivery strategy based on real-time monitoring data to achieve automated iterative optimization of advertising budget allocation, creative content, and delivery time periods.

2. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The user profile building module further includes: Cross-platform data cleaning unit: performs noise reduction, deduplication, and standardization processing on heterogeneous data from mobile applications, web logs, and e-commerce transaction systems; Real-time interest prediction submodule: Utilizes LSTM neural network to analyze user behavior sequences, dynamically updates interest tag weights, and generates an interest decay model for the next 7 days.

3. The integrated system for new media advertising placement and monitoring according to claim 2, characterized in that, The cross-platform data cleaning unit is configured with: A knowledge graph-based data association engine identifies the anonymous IDs of the same user on different platforms and establishes mapping relationships. The data quality verification rule base automatically repairs or removes outliers, missing values, and conflicting data.

4. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The intelligent advertising delivery engine includes: Dynamic Creative Generator: Automatically synthesizes advertising creatives adapted to different terminal resolutions and scenarios based on generative adversarial networks; Bidding strategy optimizer: Employs a multi-armed slot machine algorithm to adjust bidding parameters based on historical bidding success rates and real-time competition intensity.

5. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The end-to-end monitoring module further includes: Abnormal Traffic Detection Unit: Uses Gaussian mixture model to perform cluster analysis on clickstream data to identify bot traffic, proxy IPs, and click farm behavior; Cross-device attribution unit: Combining device fingerprinting technology, Wi-Fi probes, and cookie mapping, it associates the advertising interaction sequences of the same user across mobile phones, tablets, and smart TVs.

6. The integrated system for new media advertising placement and monitoring according to claim 5, characterized in that, The abnormal traffic detection unit is configured with: A click fraud detection model based on time series analysis can detect short-term high-frequency clicks and non-human operation trajectories, such as uniform-speed swiping. Traffic blacklist / whitelist database, which is synchronized in real time with malicious IPs and device IDs marked by third-party anti-fraud platforms.

7. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The data analysis and performance evaluation module includes: User lifetime value prediction model: The Cox proportional hazards regression algorithm is used to predict the long-term value of users by combining user activity and consumption frequency. Sentiment Analysis Unit: Analyzes user comments using the BERT natural language processing model to quantify the impact of advertising on brand sentiment.

8. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The closed-loop optimization control module performs the following functions: Budget dynamic allocator: Based on a linear programming algorithm, it allocates budget weights according to the real-time ROI of each channel; Intelligent alarm system: When the click-through rate or conversion rate deviates from the predicted value by more than 15%, it triggers an emergency adjustment of the advertising strategy; A / B testing framework: Automatically generates combinations of ad creative variations, such as copy, color scheme, and CTA buttons, and selects the optimal solution after parallel testing.

9. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, Also includes: Blockchain Evidence Storage Unit: Stores ad exposure, click and conversion data on the blockchain to generate an immutable performance verification certificate; Data visualization dashboard: Provides heatmaps, conversion funnels, and Sankey diagrams of user paths for real-time campaign data, supporting multi-dimensional drill-down analysis.

10. The integrated system for new media advertising placement and monitoring according to claim 1, characterized in that, The system ensures data security through the following methods: Homomorphic encryption technology is used to encrypt user privacy data, such as device ID and geolocation. A tiered data access permission mechanism is set up to restrict advertisers to only query aggregated statistical results and not to obtain original user information.