Advertisement putting method and system based on big data analysis

Abstract

The invention discloses an advertisement putting method and system based on big data analysis, and relates to the technical field of big data analysis, and the method comprises the following steps: analyzing user behavior data, and obtaining an initial feature; performing cross-domain behavior fusion according to the initial features to obtain user cross-domain features; performing click probability prediction according to the cross-domain features of the user to obtain a click probability result; performing conversion behavior analysis according to the click probability result to obtain a conversion probability result; performing audience layering calculation according to a conversion probability result to obtain an audience layering set; and constructing an advertisement display sequence according to the audience hierarchical set to obtain the advertisement display sequence. According to the method, the click probability and the cross-domain user features are combined for modeling, and the conversion probability of the user after the advertisement is clicked is calculated, so that the accuracy and the flexibility of conversion prediction are improved, a personalized putting basis can be provided for different users, and the effectiveness of advertisement putting is enhanced.

Description

Technical Field

[0001] This invention relates to the field of big data analytics technology, specifically to a method and system for advertising delivery based on big data analytics. Background Technology

[0002] With the development of the internet and mobile devices, advertising has gradually shifted from traditional static audience targeting to data-driven precision targeting. Big data analytics-based advertising methods can achieve personalized ad displays and improve ad conversion rates by collecting and analyzing user behavior data. However, most existing advertising methods still have some problems.

[0003] Current technologies have shortcomings in conversion behavior analysis: existing advertising methods mostly rely on click history or simple statistical features, ignoring the correlation between click behavior and conversion behavior, resulting in insufficient predictive flexibility and accuracy, and an inability to develop differentiated advertising strategies for users with different conversion tendencies, thus limiting the effectiveness of advertising. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a big data analysis-based advertising delivery method and system to solve the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides an advertising delivery method and system based on big data analysis, comprising the following steps: S1. Analyze user behavior data to obtain initial features; S2. Perform cross-domain behavior fusion based on the initial features to obtain user cross-domain features; S3. Predict the click probability based on the user's cross-domain characteristics to obtain the click probability result; S4. Analyze conversion behavior based on the click probability results to obtain the conversion probability results; S5. Calculate audience segmentation based on the conversion probability results to obtain the audience segmentation set; S6. Construct the ad display sequence based on the audience segmentation set to obtain the ad display sequence.

[0006] To further optimize this technical solution, the user behavior data analysis in step S1 includes: By transforming user behavior data into a unified feature dimension, we can obtain initial features that reflect users' true preferences, enhance user identifiability, and provide a solid foundation for subsequent steps.

[0007] To further optimize this technical solution, the cross-domain behavior fusion in step S2 includes: By fusing initial user characteristics and eliminating platform isolation, cross-domain user characteristics are obtained, thereby forming a complete user profile, improving the accuracy of user profiling, and utilizing users' multi-platform behavioral data in advertising.

[0008] To further optimize this technical solution, the click probability prediction in step S3 includes: Based on the cross-domain user characteristics obtained in step S2, predictive modeling is performed to model the probability of users clicking on ads, quantifying the likelihood of users clicking on ads, thereby accurately predicting user interests, reducing invalid displays, and improving ad delivery efficiency and overall delivery revenue.

[0009] To further optimize this technical solution, the conversion behavior analysis in step S4 includes: Based on the click probability results obtained in step S3, the conversion behavior analysis is performed using the conversion analysis model to calculate the conversion probability and obtain the conversion probability results.

[0010] To further optimize this technical solution, the transformation analysis model includes:

[0011] in: : Conversion probability; Click probability; Cross-domain characteristics; : Transform conditional probability.

[0012] To further optimize this technical solution, the transformation conditional probability includes:

[0013] in: Bias parameters; Feature weight parameters; : No. Cross-domain characteristics of users; The number of types of cross-domain user characteristics; The conditional probability of conversion after a click is obtained by superimposing the combined effects of various cross-domain features on user conversion behavior.

[0014] To further optimize this technical solution, the audience segmentation calculation in step S5 includes: Based on the conversion probability results, the user group is divided according to the conversion probability, which expands the advertising campaign from single-point prediction to structured group management, making it easier to control in batches, thereby improving the overall campaign effect and providing data support for subsequent steps.

[0015] To further optimize this technical solution, the construction of the advertisement display sequence in step S6 includes: By using time-series management, an ad display sequence is constructed based on audience segmentation sets. Different audience segments are mapped to the priority order and push batches of ad displays, forming an executable time-series logic that enables ads to be displayed in an orderly manner during actual pushes.

[0016] This technical solution has been further optimized, including the following functional modules: The module includes a feature acquisition module, a behavior fusion module, a probability calculation module, an audience segmentation module, and a sequence construction module.

[0017] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein the computer program instructions, when executed by the processor, implement the steps of an advertising delivery method and system for big data analysis as described in the first aspect of the present invention.

[0018] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of an advertising delivery method and system for big data analysis as described in the first aspect of the present invention.

[0019] Compared with existing technologies, the present invention provides an advertising delivery method and system based on big data analysis, which has the following beneficial effects: This big data analytics-based advertising delivery method and system combines click probability with cross-domain user characteristics through a conversion analysis model to calculate the conversion probability of users after clicking on an ad. This improves the accuracy and flexibility of conversion prediction, provides personalized delivery criteria for different users, and enhances the effectiveness of advertising delivery. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating an advertising delivery method and system based on big data analysis proposed in this invention. Figure 2 This is a flowchart illustrating the conversion analysis model of an advertising delivery method and system based on big data analysis proposed in this invention. Figure 3 This is a schematic diagram of the modules of an advertising delivery method and system based on big data analysis proposed in this invention. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments. Example 1

[0025] Reference Figures 1-2 This is the first embodiment of the present invention, which provides an advertising delivery method based on big data analysis, including the following steps: S1. Analyze user behavior data to obtain initial features.

[0026] In this embodiment, the user behavior data analysis includes: During the advertising campaign, advertisers need to understand users' interests, preferences, activity habits, and historical interactions; otherwise, the targeting will be unclear, and ads may receive ineffective impressions or have low conversion rates. User behavior data is the best source of information reflecting users' true preferences. These behaviors exhibit strong individual differences and temporal characteristics, making them the core basis for characterizing users.

[0027] The purpose of this step is to transform the raw, complex user behavior data into unified feature dimensions, such as click frequency, interest tags, and active time periods, to obtain initial features that reflect users' true preferences and can be used for calculation and analysis. This enhances user identifiability and provides a solid foundation for subsequent steps.

[0028] The implementation methods for this step include: Determine data sources: Select behavior-related data fields from the historical logs of the advertising platform, including ad click records (reflecting user interests), browsing dwell time (reflecting user attention to content), interaction events, device identification, access timestamps, etc., and select the basic fields that can truly characterize user behavior to provide input sources for subsequent steps; Data structuring: Hadoop distributed computing technology is used to process data in a structured manner. Hadoop can divide the raw log file into parallel processing task units, which can run simultaneously in a multi-node environment, thereby significantly shortening the parsing time and ensuring that large-scale data can be parsed within a reasonable time. This transforms the data from an unordered state into regular entries, providing a neat data format. Feature field extraction: Select the most representative feature fields of user behavior from structured data, such as daily click count (describes user activity level), ad interaction depth (reflects user engagement), average dwell time (distinguishes between shallow browsing and deep reading), interest category tags (reflects user content preferences), and commonly used device types (reveal user usage habits), etc. Data cleaning and summarization: Deleting or correcting abnormal records, removing data records with serious missing values ​​or abnormalities, such as unreasonable excessively long dwell time or excessively frequent click events. After cleaning and summarization, the feature fields can better reflect the user's real behavior patterns and improve the credibility of the data. Generate an initial feature set: The cleaned features are uniformly encoded and quantified to give them a unified format, such as numerical features (number of clicks), categorical features (interest tags), and time features (active time periods), so as to comprehensively and systematically describe the user's basic behavioral characteristics. Storage and Output: The generated initial feature set is indexed and stored according to user ID to ensure that each user corresponds to a unique feature entry. Distributed databases are usually used for storage to support fast reading and retrieval of large-scale user data and provide data input for subsequent steps.

[0029] S2. Perform cross-domain behavior fusion based on the initial features to obtain user cross-domain features.

[0030] In this embodiment, the cross-domain behavior fusion includes: In the real-world advertising environment, user behavior is not limited to a single platform. For example, user purchasing behavior on e-commerce platforms, interactive behavior on social media platforms, and browsing behavior on news platforms can all reflect different aspects of interests and preferences. Relying solely on the initial characteristics of a single platform can easily create "information silos," leading to one-sided user profiles that fail to accurately reflect users' true needs, thereby reducing the effectiveness of advertising.

[0031] The purpose of this step is to merge the initial characteristics of users from different sources and scenarios, eliminate platform isolation, and obtain cross-domain user characteristics at the cross-domain level. This enables advertising to utilize users' multi-platform behavioral data to form a complete profile and improve the accuracy of user profiling.

[0032] The implementation methods for this step include: Establish user identifier mapping: First, determine a unified identifier for user identity across platforms, such as using an encrypted UID as a cross-domain index. Under the premise of ensuring privacy and security, accurately match behavioral data from e-commerce, social, news and other platforms to the same user, and ensure that the same user can be correctly associated on different platforms. Feature Dimension Alignment: Due to the differences in feature fields across different platforms, such as product categories and browsing categories on e-commerce platforms, interest topics and interaction frequency on social media platforms, and browsing time periods and content types on news platforms, dimension mapping is required to map the features of different platforms to a unified standardized dimension, thereby enabling comparison and fusion within the same semantic framework. For example, "product categories" on e-commerce platforms can be mapped to interest categories, "interaction topics" on social media platforms can be transformed into interest tags, and "browsing time periods" on news platforms can be normalized to active time periods. Feature fusion calculation: After dimensional alignment is completed, features from multiple platforms are integrated, and features of the same category are weighted and merged. For example, for interest category features, weights can be assigned according to the intensity of user behavior on each platform (such as number of clicks, interaction depth, and purchase frequency), and then weighted and merged into a unified interest preference value. This integrates multi-source behavioral information into a quantifiable and analyzable comprehensive feature, enabling user behavior on different platforms to complement each other, improve feature completeness and information content, and better reflect users' true preferences and potential needs. Conflict handling and completion: When feature values ​​conflict between different platforms, such as one platform showing user preference for "technology" and another showing preference for "entertainment", the weight allocation is determined by comparing rules such as behavior frequency, time relevance, and platform weight, and the final comprehensive preference value is determined. For missing features, such as when a user has no interest record on a certain platform, historical data or similar user behavior is used to reasonably complete the feature, thereby ensuring the integrity and consistency of cross-domain features and avoiding the impact of data conflict or missing data on the accuracy of subsequent modeling. Generate cross-domain feature set: Store all fused results as user cross-domain feature set, with each user corresponding to a complete entry, which contains comprehensive attributes from different platforms, such as comprehensive interest categories, cross-platform active time periods, cross-domain consumption tendencies, etc., to provide multi-dimensional and panoramic user information for subsequent steps.

[0033] S3. Predict the click probability based on the user's cross-domain characteristics to obtain the click probability result.

[0034] In this embodiment, the click probability prediction includes: In advertising scenarios, relying solely on single behavioral data for click prediction can easily lead to biases. For example, a user may frequently browse certain topics on social media platforms but never purchase related products on e-commerce platforms. Relying solely on social behavior will overestimate the probability of clicks.

[0035] The purpose of this step is to predict and model the probability of users clicking on ads based on the cross-domain characteristics of users obtained in step S2, quantify the likelihood of users clicking on ads, thereby accurately predicting user interests, reducing invalid impressions, and improving ad delivery efficiency and overall ad delivery revenue.

[0036] The implementation methods for this step include: The mature Logistic Regression model is used for click probability modeling. This model maps the linear output to the (0,1) interval using the Sigmoid function, making it suitable for predicting click probability and computationally efficient. The steps include: Feature input preparation: Input the user cross-domain features generated in step S2, including multi-dimensional features such as interest preferences, active time periods, and consumption tendencies, to ensure that the data information used by the model during training and prediction is more comprehensive, thereby avoiding prediction bias; Numericalization of features: Cross-domain features contain a large amount of categorical data, such as interest categories that may be "technology", "entertainment", and "food", and consumption preferences that may be "price-sensitive" and "brand-oriented". The model cannot directly identify these features and needs to convert them into numerical forms. For example, one-hot encoding (which converts each category into a binary variable, e.g., if the user's interest is "technology", then technology is 1 and the other categories are 0) can be used to process interest category features, enabling the model to identify and calculate the impact of different features. Model training: Using historical ad display and click data as training samples, a logistic regression model is trained. During the training process, the model continuously adjusts the weight coefficients of each feature according to the relationship between user cross-domain features and click results, so as to fit the pattern between user features and click behavior and make the predicted probability value as close as possible to the actual click result. Probability Calculation: After training, the logistic regression model has the ability to predict new input data. When a new advertisement needs to be displayed, the cross-domain features of the target user are extracted and input into the model. The logistic regression model will output a probability value between 0 and 1, which represents the likelihood of the user clicking the advertisement, facilitating subsequent decision-making. Output Results: The probability results are presented as structured output, including the probability of users clicking on the ads. This information is provided for subsequent steps and serves as a key basis for further evaluation and improvement of ad performance.

[0037] S4. Analyze conversion behavior based on the click probability results to obtain conversion probability results.

[0038] In this embodiment, the conversion behavior analysis includes: In advertising scenarios, a user clicking on an ad is merely an intermediate action that can only partially reflect the likelihood of conversion and cannot be directly used for ad placement decisions. What advertisers are truly concerned with are "conversion behaviors" (such as purchases, registrations, or downloads).

[0039] The purpose of this step is to use a conversion analysis model to analyze conversion behavior based on the click probability results obtained in step S3, calculate the conversion probability, and obtain the conversion probability results. This provides a more commercially valuable evaluation indicator for advertising placement and avoids misjudgments caused by relying solely on click-through rate.

[0040] Furthermore, the transformation analysis model includes:

[0041] in: Conversion probability represents the probability that a user will convert from a target ad. It ranges from 0 to 1, with a higher value indicating a greater likelihood of conversion. Click probability represents the likelihood of a user clicking on the target ad. It is obtained from step S3 and is used to calculate the conversion probability in conjunction with the conversion conditional probability after the click. Cross-domain features represent the set of user cross-domain features obtained in step S2, including interest preferences, cross-platform active time periods, consumption tendencies, etc., which describe the user's behavioral tendency to complete conversion after clicking. The values ​​of continuous features (such as active time periods and purchase amount preferences) can be taken from 0 to 1 according to the actual normalization results of the features. The values ​​of categorical features (such as interest categories) can be converted into binary values ​​of 0 or 1 through one-hot encoding. : Conversion conditional probability, calculated using a conversion probability modeling function, describes the likelihood of a user completing a conversion after clicking.

[0042] Furthermore, the transition conditional probability includes:

[0043] in: Bias parameter, as the intercept term, is used to adjust the baseline value to ensure that a reasonable basic conversion probability can still be obtained when all user features are 0. It is obtained through maximum likelihood estimation during the training phase. Feature weight parameters reflect the positive or negative impact of different features on user conversion behavior. For example, features with strong purchase intentions correspond to positive weights, while irrelevant features have weights close to zero or negative. These weights are adjusted based on historical click and conversion data. : No. Cross-domain characteristics of users; The number of types of cross-domain user characteristics; The conditional probability of conversion after a click is obtained by superimposing the combined effects of various cross-domain features on user conversion behavior.

[0044] This model describes how to calculate the conversion conditional probability after a click based on the user's cross-domain characteristics, and then combine it with the user's click probability to obtain the final conversion probability.

[0045] Traditional conversion prediction methods often use single-layer logistic regression or single-platform features for prediction, lacking hierarchical relationship modeling between clicks and conversions, resulting in insufficient prediction flexibility and accuracy. In contrast, this model adopts hierarchical probability modeling, combining click probability with cross-domain features as input, which is logically more in line with user behavior paths. Furthermore, by introducing cross-domain features as conversion influencing factors, it solves the problem of insufficient information in traditional models, improves the accuracy and flexibility of conversion probability prediction, and the output results directly reflect the commercial value of advertising, providing a basis for personalized targeting, improving the reliability of resource allocation indicators, and enhancing the effectiveness of advertising.

[0046] The steps for using this model include: Data Acquisition: Obtain user cross-domain characteristics from step S2. The click probability result is obtained from step S3. ; Hierarchical modeling: using cross-domain features as auxiliary factors By learning from historical ad delivery data, the system fits the "click-to-conversion" behavior pattern and calculates the conditional probability of conversion after a click. ; Conversion probability calculation: based on click probability Combining the transformation conditional probability Calculate the final conversion probability. This provides more commercially valuable evaluation metrics for advertising placement.

[0047] S5. Calculate audience segmentation based on the conversion probability results to obtain the audience segmentation set.

[0048] In this embodiment, the audience segmentation calculation includes: In step S4, the conversion probability results of each user after clicking the ad are obtained, which can reflect the conversion probability of an individual user. However, it lacks a hierarchical structure expression of the audience. Advertisers usually need to formulate different advertising budgets and resource allocations based on different levels of audience groups. Therefore, users must be orderly segmented according to conversion probability.

[0049] The purpose of this step is to divide the user group into several levels according to the conversion probability results, such as high-value audience, medium-value audience and low-value audience, so that the advertising campaign can be expanded from single-point prediction to group structured management, which is convenient for batch control, thereby improving the overall campaign effect and providing data support for subsequent steps.

[0050] The implementation methods for this step include: Calculate user probability distribution characteristics: Collect the conversion probability results of all users, construct the group conversion probability distribution, and obtain the interval characteristics of the overall audience conversion probability by calculating statistical indicators such as the mean (representing the overall level), median (representing the central position of the group), and quantiles (representing different levels of the group distribution). For example, whether most users are concentrated in the low probability interval or the distribution is relatively balanced, this provides a quantitative basis for setting subsequent stratification boundaries, making the stratification more scientific and interpretable. Set stratification boundaries: Based on the distribution statistics, divide the conversion probability interval into multiple segments, such as low probability interval [0, 0.3), medium probability interval [0.3, 0.6], and high probability interval [0.6, 1]. Each segment corresponds to an audience level. The interval with higher values ​​represents a higher conversion tendency. If most users are concentrated in the low value interval, the interval should be subdivided to further stratify the low probability group. If users are relatively sparse in a certain interval, the interval range can be expanded to avoid creating too few levels, ensure that the number of users in each level is reasonable, avoid extreme imbalance between levels, and ensure that the stratified set has stability and representativeness. Perform user tier classification calculation: Iterate through the conversion probability results of each user, compare them with the boundary, classify them into the corresponding tier interval, establish a user-tier correspondence table, and form a preliminary tier set; Generate audience segmentation sets: Aggregate the tiered classification results of all users to obtain high conversion-prone audience sets, medium conversion-prone audience sets, and low conversion-prone audience sets. Each set retains individual user information, while the set as a whole has a hierarchical identifier, which allows advertising to allocate resources and evaluate performance at the set level. Output structured stratified results: Organize the above aggregated audience sets into structured results, output audience stratification sets, clearly identify the scope, number of users and conversion tendency of each stratification set, so as to retain the conversion probability of individual users and form a stratified structure at the group level, providing a direct basis for subsequent advertising placement and resource allocation.

[0051] S6. Construct the ad display sequence based on the audience segmentation set to obtain the ad display sequence.

[0052] In this embodiment, the construction of the advertisement display sequence includes: Most existing ad delivery methods rely solely on static audience segments for ad pushes, lacking a sequential design that leads to a fragmented delivery process. Effective ad delivery requires not only identifying the target audience but also arranging the ad display order; otherwise, it can easily result in ad duplication, audience fatigue, or failure to prioritize high-value users.

[0053] The purpose of this step is to construct an ad display sequence based on audience segmentation through time-series management. This maps different audience segments to the priority order and push batches of ad displays, forming an executable time-series logic that enables ads to be displayed in an orderly manner during actual pushes, thereby balancing accuracy and breadth and avoiding resource waste.

[0054] Furthermore, the time-series management includes: Determine display priority: Based on the audience segmentation set generated in step S5, assign different display priorities to different segments. Generally, the high conversion-prone audience segment is given the highest priority, followed by the medium conversion-prone audience segment, and the low conversion-prone audience segment is given the lowest priority. This can be adjusted based on the user's historical ad response data. Construct display batches: Divide the user set of each level into several display batches. Each batch corresponds to a delivery time window for displaying ads in stages, avoiding resource waste or user fatigue caused by concentrated push. For example, high-level user groups can be concentrated in the early batch to quickly capture conversion opportunities, mid-level user groups can be distributed in the mid-term batch to avoid resource squeezing, and low-level user groups can be arranged in the later batch to form supplementary exposure. Set display intervals and frequencies: Set display intervals for different batches in the sequence. For example, the display interval for ads of high-conversion groups is shorter to ensure high-frequency reach, the medium interval is used for mid-level groups to avoid overexposure, and the long interval is used for low-level groups to save resources. This keeps the display frequency within a reasonable range so that users can maintain the impression of the ads without being averse to them due to excessive repetition. Create an ad display sequence: Combine audience priority, display batches, intervals and frequencies to generate an ad display sequence, specifying when, to which user groups, and at what frequency to display ads, making ad delivery more efficient and precise; Output execution structure: Organize the ad display sequence into a structured result, including hierarchical information, batch time, display interval and frequency, and output the final ad display sequence as input for campaign execution. Example 2

[0055] Reference Figure 3 This is the second embodiment of the present invention, which provides an advertising delivery system based on big data analysis, including the following functional modules: Feature acquisition module: Collects user behavior data (such as browsing history, click history, transaction records, etc.) and extracts initial features to provide input data for cross-domain analysis; Behavior fusion module: Fusion of user behavior features from different sources to construct cross-domain user features, realize a panoramic profile of user features, and enhance the expressive power of user features; Probability Calculation Module: Based on the fused cross-domain features, a click probability prediction model is established to output the ad click probability for each user. A conversion analysis model is also established to calculate the conversion probability of users after clicking on ads, providing a basis for audience segmentation. Audience segmentation module: Based on user conversion probability, the user group is divided into high, medium and low conversion tendency levels, providing a structured audience for advertising. Sequence building module: Generates ad display sequences based on audience segmentation sets, including hierarchical priority, display batches, time intervals, and display frequency. Example 3

[0056] This embodiment also provides a computer device applicable to a big data analysis-based advertising delivery method and system, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the big data analysis-based advertising delivery method and system proposed in the above embodiment.

[0057] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements an advertising delivery method and system based on big data analysis as described in the above embodiments.

[0058] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0059] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0060] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0061] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0062] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0063] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A big data analytics-based advertising delivery method, characterized in that, Includes the following steps: S1. Analyze user behavior data to obtain initial features; S2. Perform cross-domain behavior fusion based on the initial features to obtain user cross-domain features; S3. Predict the click probability based on the user's cross-domain characteristics to obtain the click probability result; S4. Analyze conversion behavior based on the click probability results to obtain the conversion probability results; S5. Calculate audience segmentation based on the conversion probability results to obtain the audience segmentation set; S6. Construct the ad display sequence based on the audience segmentation set to obtain the ad display sequence.

2. The advertising placement method based on big data analysis according to claim 1, characterized in that, The user behavior data analysis in step S1 includes: By transforming user behavior data into a unified feature dimension, we can obtain initial features that reflect users' true preferences, enhance user identifiability, and provide a solid foundation for subsequent steps.

3. The advertising placement method based on big data analysis according to claim 1, characterized in that, The cross-domain behavior fusion in step S2 includes: By fusing initial user characteristics and eliminating platform isolation, cross-domain user characteristics are obtained, thereby forming a complete user profile, improving the accuracy of user profiling, and utilizing users' multi-platform behavioral data in advertising.

4. The advertising placement method based on big data analysis according to claim 1, characterized in that, The click probability prediction in step S3 includes: Based on the cross-domain user characteristics obtained in step S2, predictive modeling is performed to model the probability of users clicking on ads, quantifying the likelihood of users clicking on ads, thereby accurately predicting user interests, reducing invalid displays, and improving ad delivery efficiency and overall delivery revenue.

5. The advertising placement method based on big data analysis according to claim 1, characterized in that, The conversion behavior analysis in step S4 includes: Based on the click probability results obtained in step S3, the conversion behavior analysis is performed using the conversion analysis model to calculate the conversion probability and obtain the conversion probability results.

6. The advertising placement method based on big data analysis according to claim 5, characterized in that, The transformation analysis model includes: , in: : Conversion probability; Click probability; Cross-domain characteristics; : Transform conditional probability.

7. The advertising placement method based on big data analysis according to claim 6, characterized in that, The transition conditional probabilities include: ， in: Bias parameters; Feature weight parameters; : No. Cross-domain characteristics of users; The number of types of cross-domain user characteristics; The conditional probability of conversion after a click is obtained by superimposing the combined effects of various cross-domain features on user conversion behavior.

8. The advertising placement method based on big data analysis according to claim 1, characterized in that, The audience segmentation calculation in step S5 includes: Based on the conversion probability results, the user group is divided according to the conversion probability, which expands the advertising campaign from single-point prediction to structured group management, making it easier to control in batches, thereby improving the overall campaign effect and providing data support for subsequent steps.

9. The advertising placement method based on big data analysis according to claim 1, characterized in that, The construction of the advertisement display sequence in step S6 includes: By using time-series management, an ad display sequence is constructed based on audience segmentation sets. Different audience segments are mapped to the priority order and push batch of ad displays, forming an executable time-series logic that enables ads to be displayed in an orderly manner during actual push.

10. The advertising delivery system based on big data analysis according to claim 1, characterized in that, Includes the following functional modules: The module includes a feature acquisition module, a behavior fusion module, a probability calculation module, an audience segmentation module, and a sequence construction module.

Images