Marketing data analysis method and system
By adopting real-time data flow acquisition and parallel processing, multi-dimensional user classification and network attack protection in the marketing data analysis system, the problems of insufficient data classification, slow collection speed and weak security are solved, and efficient and accurate data processing and Safety protection.
Patent Information
- Application Number
- CN202411124944.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-08-16
AI Technical Summary
The existing marketing data analysis system has problems such as insufficient data classification, slow data collection speed and weak data security protection, which affects the effectiveness and security of data processing.
By adopting real-time data flow acquisition technology and parallel processing framework in the data acquisition module, data acquisition efficiency and processing speed are improved; in the user classification module, multi-dimensional user attributes and behavior characteristics are considered for classification; in the data security module, monitoring is set up Threads monitor network ports, identify and prevent cyber attacks.
It realizes the rapid and effective processing of massive marketing data, improves user classification accuracy and data collection efficiency, and enhances data security, ensuring the real-time and security of the system.
Smart Images

Figure CN119025729A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present invention relate to data processing technology, and in particular to a marketing data analysis method and system. Background Art
[0002] With the rapid development of big data technology, enterprises are increasingly relying on data analysis systems to guide market decisions and improve competitiveness. At present, there are many marketing data analysis systems based on big data on the market. These systems usually include collection modules, databases, data analysis modules, intelligent screening modules and data usage modules.
[0003] For example, in some existing technical solutions, a marketing data analysis system based on big data can collect review data of all merchants in the mobile network through a data collection module, and analyze the data through subsequent modules to screen out false marketing behaviors and identify user needs.
[0004] However, in actual applications, such systems often have insufficient data classification, which results in the system being unable to effectively use the data for in-depth analysis, thus affecting the formulation of marketing strategies. In addition, the slow data collection speed makes the data less real-time and the system's defense against network attacks weak, thus affecting the effectiveness and security of data processing.
[0005] Therefore, it is necessary to propose an improved solution to these problems, which can more effectively process massive data, improve classification accuracy, and enhance data security and collection efficiency. Summary of the invention
[0006] The embodiments of the present invention provide a marketing data analysis method and system for processing massive marketing data more quickly and effectively and realizing accurate classification of users.
[0007] In a first aspect, an embodiment of the present invention provides a marketing data analysis system, including:
[0008] The data collection module is used to collect marketing data from real-time data sources, pre-process the marketing data, and store the pre-processed marketing data in the database;
[0009] The data analysis module is used to obtain marketing data from the database and analyze and identify the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result of the authenticity of the marketing data;
[0010] The user classification module is used to obtain marketing data from the database, classify users according to a preset classification model to obtain classification results, and determine the push strategy for different categories of users based on the set push strategy and classification results.
[0011] In the above scheme, the data collection module collects marketing data from the real-time data source to achieve efficient collection of marketing data. After the marketing data is collected, the marketing data is preprocessed in parallel in an asynchronous manner, which can increase the processing speed of the marketing data. Next, the data analysis module can obtain the preprocessed marketing data from the database, and identify the authenticity of the marketing data according to the preset recognition model analysis, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior based on the recognition result of the authenticity of the marketing data. When it is determined that the merchant corresponding to the marketing data has false marketing behavior, the merchant can be marked to reduce the frequency of recommendation for the merchant. At the same time, the user classification module can classify users according to the marketing data obtained from the database using the preset classification model to obtain accurate user categories, so as to achieve accurate push for users of different categories according to the classification results and the set push strategy.
[0012] In a possible design of the first aspect, the data acquisition module includes:
[0013] A real-time data collection unit, used to collect marketing data from a server corresponding to a real-time data source, where the real-time data source includes at least social media, an e-commerce platform, and a real-time log stream;
[0014] The asynchronous data processing unit is used to asynchronously pre-process the marketing data collected by the real-time data collection unit and store the pre-processed marketing data in the database.
[0015] In the above scheme, by collecting marketing data from the server corresponding to the real-time data source through the real-time data collection unit, marketing data reflecting the latest market dynamics and consumer behaviors can be obtained; by parallel processing of the collected marketing data through the asynchronous data processing unit, the data processing speed can be improved.
[0016] In a possible design of the first aspect, the real-time data collection unit is specifically used to: subscribe to marketing data of social media, marketing data of e-commerce platforms, and marketing data of log streams from a Kafka server; loop through messages received from the Kafka server and extract marketing data from the messages; and send the extracted marketing data to an asynchronous data processing unit for processing.
[0017] In the above scheme, considering that the Kafka server is a distributed stream processing platform with the characteristics of high performance, persistence, distributed structure, publish-subscribe model, fault tolerance and horizontal scalability, when collecting marketing data from real-time data sources, the embodiment of the present invention subscribes to the marketing data of each real-time data source from the Kafka server to ensure the real-time and diversity of the collected marketing data, thereby achieving more accurate user classification and false behavior judgment.
[0018] In a possible design of the first aspect, the asynchronous data processing unit is specifically used to: receive marketing data from the real-time data collection unit, perform data cleaning and data format conversion on the marketing data through an asynchronous task function, and store the processed marketing data in a database through a data saving function; wherein data cleaning and data format conversion belong to preprocessing.
[0019] In the above scheme, a method for processing marketing data through an asynchronous data processing unit is provided, which can process the marketing data in parallel during the data collection process and improve the data processing speed.
[0020] In a possible design of the first aspect, the data analysis module analyzes and identifies the authenticity of the marketing data according to a preset identification model, including:
[0021] Extract text features and numerical features of marketing data through preset recognition models;
[0022] The word frequency-inverse document frequency TF-IDF value and sentiment score corresponding to the text features are calculated through the preset recognition model. The text length of the marketing data is calculated based on the numerical features and the timestamp features are extracted. In addition, the product categories and user levels in the marketing data are encoded to obtain category features.
[0023] Obtain feature matrix based on text features, numerical features, and category features;
[0024] Based on the feature matrix, TF-IDF value and sentiment score, the authenticity of the marketing data is identified.
[0025] In the above scheme, by extracting the features of the marketing data, calculating the TF-IDF value and the sentiment score, the authenticity recognition result of the marketing data is obtained, so that the system can determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result, thereby screening out merchants with false marketing behavior.
[0026] In a possible design of the first aspect, the user classification module obtains marketing data from a database, and classifies users according to a preset classification model to obtain classification results including:
[0027] Extracting user purchase record feature data from marketing data; the user purchase record feature data includes user tag features, purchase frequency features, purchase amount features and geographic location features;
[0028] Standardize the user purchase record feature data to obtain the user data to be classified;
[0029] The user data to be classified is clustered and classified through the preset classification model to obtain the classification result.
[0030] In the above scheme, the user classification module can extract user purchase record feature data from the marketing data, and obtain classification results about the user based on the user purchase record feature data, thereby realizing multi-dimensional classification of users, which helps to understand the user group more carefully and thus achieve more accurate push.
[0031] In a possible design of the first aspect, the user classification module performs standardization processing on the user purchase record feature data to obtain the user data to be classified, including:
[0032] By formula Standardize the user purchase record feature data to obtain the user data to be classified;
[0033] Among them, z represents the user data to be classified, x represents the user purchase record feature data, μ represents the mean of each feature in the user purchase record feature data, and σ represents the standard deviation of each feature in the user purchase record feature data.
[0034] In the above scheme, the user purchase record feature data can be standardized to make the numerical ranges and scales of different features more consistent, thereby avoiding that some features have a large numerical range and affect the reliability of the preset classification model.
[0035] In a possible design of the first aspect, the system also includes: a data security module, which is used to issue an alarm message when a network attack behavior is identified, and record a network attack log to facilitate security audits.
[0036] In the above scheme, data security protection is achieved through the data security module, and it can issue an alarm message when a network attack behavior is identified, and record the network attack log, such as the time, type, source and other information of the network attack, so that relevant staff can take necessary measures in time to find potential security loopholes.
[0037] In a possible design of the first aspect, the data security module is specifically used to: according to the set listening address and port, start the monitoring thread to monitor the data of the network port corresponding to the listening address and port, and obtain the monitoring data;
[0038] When a preset network attack behavior is determined to occur based on the monitoring data, an alarm message is issued, a network attack log is recorded, and an interrupt signal is sent to disconnect the network connection.
[0039] In the above scheme, the data security module starts a monitoring thread to listen to the set listening address and port to obtain the listening data, so that it can determine whether a preset network attack behavior is currently occurring based on the listening data, and disconnect the network connection in time through an interrupt signal when it is determined that a preset network attack behavior has occurred to prevent the attack from spreading.
[0040] In a second aspect, an embodiment of the present invention provides a marketing data analysis method, which is applied to a marketing data analysis system as described in any one of the first aspects of the embodiments of the present invention, and the method includes:
[0041] The data collection module in the marketing data analysis system collects marketing data from real-time data sources;
[0042] The data collection module pre-processes the marketing data and stores the pre-processed marketing data in the database;
[0043] The data analysis module in the marketing data analysis system obtains marketing data from the database, and analyzes and identifies the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result of the authenticity of the marketing data;
[0044] The user classification module in the marketing data analysis system obtains marketing data from the database, classifies users according to a preset classification model to obtain classification results, and determines the push strategy for different categories of users based on the set push strategy and classification results.
[0045] In the above scheme, the data collection module collects marketing data from the real-time data source, pre-processes the marketing data, and stores the pre-processed marketing data in the database. The data analysis module obtains the marketing data from the database, and analyzes and identifies the authenticity of the marketing data according to the preset recognition model, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior according to the recognition result. The user classification module obtains the marketing data from the database, classifies the users according to the preset classification model to obtain the classification results, and determines the push strategy for different categories of users based on the set push strategy and classification results. In this way, the massive marketing data can be processed more quickly and effectively, and the accurate classification of users can be achieved.
[0046] In a third aspect, an embodiment of the present invention provides an electronic device, including:
[0047] processor; and,
[0048] A memory for storing executable instructions of the processor;
[0049] Wherein, the processor is configured to perform the method in the second aspect by executing executable instructions.
[0050] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the method in the second aspect. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present invention and, together with the description, serve to explain the principles of the embodiments of the present invention.
[0052] Figure 1 A schematic diagram of the structure of a marketing data analysis system provided by an embodiment of the present invention is shown;
[0053] Figure 2 A schematic diagram showing the structure of another marketing data analysis system according to an embodiment of the present invention is shown;
[0054] Figure 3 A schematic diagram showing the structure of another marketing data analysis system according to an embodiment of the present invention is shown;
[0055] Figure 4 A schematic diagram showing a flow chart of a marketing data analysis method provided by an embodiment of the present invention;
[0056] Figure 5 It is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present invention.
[0057] The above drawings have shown specific embodiments of the present invention, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the embodiments of the present invention in any way, but to illustrate the concepts of the embodiments of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION
[0058] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the embodiments of the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the embodiments of the present invention as detailed in the appended claims.
[0059] Marketing data analysis is a method used by companies to evaluate the effectiveness of their marketing strategies and activities. In marketing data analysis, marketing data usually refers to various data related to marketing, advertising, sales, and customer interaction. This data can include but is not limited to:
[0060] Sales data (including sales volume, sales channels, etc.); user data (including user demographic information, purchase history, user satisfaction, etc.); market data (including market size, market share, competitor analysis, etc.); advertising data (including advertising expenditure, click-through rate, conversion rate and advertising channels, etc.); social media data (including interactions on social media, fan growth and brand mentions, etc.); website data (including visits, user behavior, page views and bounce rate, etc.); user feedback (including user opinions collected through surveys, comments or feedback forms); promotional activity data (including promotional activity participation, promotional effectiveness, etc.); product data (including product life cycle, product portfolio and product characteristics, etc.); pricing data (including sales performance at different price points, price sensitivity, etc.).
[0061] By analyzing these data, companies can better understand market trends, consumer behavior and the effectiveness of marketing activities, thereby optimizing marketing strategies and improving return on investment.
[0062] With the rapid development of big data technology, enterprises are increasingly relying on marketing data analysis systems to guide market decisions and improve competitiveness. In order to meet this demand, a variety of marketing data analysis systems based on big data have appeared on the market. These marketing data analysis systems usually include data collection modules, databases, data analysis modules, intelligent screening modules, and data usage modules. Although these systems have improved the data processing capabilities and decision-making efficiency of enterprises to a certain extent, there are still some challenges in practical applications, such as insufficient data classification, slow data collection speed, and weak data security protection.
[0063] Specifically, existing marketing data analysis systems often only focus on data classification in a single dimension, such as classification based only on users' purchase records, while ignoring other important user attributes and behavioral characteristics. This results in the inability of marketing data analysis systems to effectively use these data for in-depth analysis, thus affecting the formulation of marketing strategies.
[0064] In addition, due to the huge amount of marketing data and the scattered data sources, the data collection methods used in traditional marketing data analysis systems often find it difficult to achieve real-time data collection, which makes the real-time performance of the collected marketing data poor and reduces the response speed and flexibility of the marketing data analysis system.
[0065] For example, with the continuous evolution of network attack methods, the existing marketing data analysis system has many loopholes in data security protection, which may not only lead to the leakage of marketing data and related analysis results data, but also affect the effectiveness and security of marketing data processing.
[0066] Therefore, in order to solve the above problems, it is necessary to propose an improved marketing data analysis system to more effectively process massive marketing data, improve user classification accuracy, and enhance data security and collection efficiency.
[0067] The embodiments of the present invention aim to solve the problems existing in the above-mentioned prior art in the marketing data analysis system, such as insufficient data classification, slow data collection speed, weak data security protection, etc.
[0068] Specifically, the concepts of the embodiments of the present invention are centered around the following core points:
[0069] Improve the efficiency of marketing data collection: Use real-time data stream collection technology to achieve real-time data collection to ensure the real-time and integrity of marketing data, and use a parallel processing framework to improve the speed and efficiency of marketing data processing.
[0070] Achieve multi-dimensional user classification: When classifying users, in addition to considering purchase records, user attributes such as age, gender, geographic location, hobbies, purchase frequency, and purchase amount are also considered to get a more comprehensive understanding of users. At the same time, by analyzing user browsing records, search history, click-through rate and other behavioral data, we can identify user preferences and behavior patterns; in addition, we can also use natural language processing technology to perform sentiment analysis on user comments to understand user satisfaction and reasons for dissatisfaction, thereby achieving multi-dimensional classification of users, which helps to understand the user group more carefully and achieve more accurate push.
[0071] Enhance data security: By setting up the data security module, the monitoring thread is started to monitor the set listening address and port to obtain the monitoring data. In this way, it is possible to determine whether a preset network attack behavior is currently occurring based on the monitoring data, and when it is determined that a preset network attack behavior has occurred, the network connection is promptly disconnected through an interrupt signal to prevent the attack from spreading.
[0072] In summary, the concept of the embodiment of the present invention is to improve the accuracy of user classification, the efficiency of marketing data collection, and enhance data security. The data collection module collects marketing data from the real-time data source to achieve efficient collection of marketing data, and after collecting the marketing data, the marketing data is preprocessed in parallel in an asynchronous manner, which can increase the processing speed of the marketing data. Then, the data analysis module can obtain the preprocessed marketing data from the database, and identify the authenticity of the marketing data according to the preset recognition model analysis, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior according to the recognition result of the authenticity of the marketing data, and when it is determined that the merchant corresponding to the marketing data has false marketing behavior, the merchant can be marked to reduce the frequency of recommendation to the merchant. At the same time, the user classification module can classify users according to the marketing data obtained from the database using the preset classification model to obtain accurate user categories, so as to achieve accurate push for users of different categories according to the classification results and the set push strategy.
[0073] The marketing data analysis method and system provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0074] Figure 1 FIG. 1 is a schematic diagram showing the structure of a marketing data analysis system provided by an embodiment of the present invention. Figure 1 As shown, the marketing data analysis system 100 provided by the embodiment of the present invention may include: a data collection module 110 , a data analysis module 120 and a user classification module 130 .
[0075] Specifically, the data collection module 110 is used to collect marketing data from a real-time data source, pre-process the marketing data, and store the pre-processed marketing data in a database.
[0076] The data analysis module 120 is used to obtain marketing data from the database and analyze and identify the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result of the authenticity of the marketing data.
[0077] The user classification module 130 is used to obtain marketing data from the database, classify users according to a preset classification model to obtain classification results, and determine push strategies for different categories of users based on the set push strategies and classification results.
[0078] Among them, the preprocessing of marketing data can include data cleaning processing and data format conversion processing. Data cleaning processing is to remove noise data in marketing data, and data format conversion processing can standardize and normalize marketing data to improve the accuracy of subsequent processing.
[0079] After preprocessing the marketing data, the data collection module 110 stores the preprocessed marketing data in a database. In one implementation, the database may be a database implemented based on a distributed storage architecture, and the distributed storage structure may be implemented based on data compression and indexing technology to optimize storage space and accelerate data retrieval, thereby achieving high-speed storage and retrieval of marketing data. In actual applications, the implementation of the database may be determined according to actual usage requirements, and this embodiment does not specifically limit this.
[0080] The preset recognition model used by the data analysis module 120 is obtained by training based on the training sample marketing data and the user behavior corresponding to the training sample marketing data, and can analyze and identify the authenticity of the marketing data to obtain an identification result about the authenticity. For example, if the marketing data is identified as false data, it means that the merchant or user corresponding to the marketing data has engaged in false marketing behavior. If the marketing data is identified as true data, it means that the merchant or user corresponding to the marketing data has not engaged in false marketing behavior.
[0081] In a possible implementation, the authenticity identification result of the marketing data may be an authenticity value. If the authenticity value of the marketing data is greater than the authenticity threshold, it means that the merchant or user corresponding to the marketing data does not engage in false marketing; if the authenticity value of the marketing data is less than or equal to the authenticity threshold, it means that the merchant or user corresponding to the marketing data engages in false marketing.
[0082] For example, assuming that the authenticity threshold is 60, the authenticity value in the authenticity identification result of marketing data A is 50, and the authenticity value in the authenticity identification result of marketing data B is 90. Since the authenticity value 50 of marketing data A is less than the authenticity threshold 60, it is determined that the merchant or user corresponding to marketing data A has engaged in false marketing behavior. Since the authenticity value 90 of marketing data B is greater than the authenticity threshold 60, it is determined that the merchant or user corresponding to marketing data B has not engaged in false marketing behavior.
[0083] The preset classification model used by the user classification module 130 is obtained by clustering the purchase record feature data of the training sample users in advance. In actual use, after obtaining the marketing data, the user classification module 130 can input the user purchase record feature data in the marketing data into the preset classification model to obtain the corresponding classification result. Afterwards, the push strategy for different categories of users can be determined based on the classification result and the set push strategy, thereby realizing personalized recommendations for different categories of users.
[0084] It should be noted that the preset classification model can realize multi-dimensional classification of users so as to more accurately understand users' preferences and behavior patterns, and thus provide more personalized services.
[0085] Exemplarily, the dimensions for classifying users may include: classification based on purchase behavior, classification based on user attributes, classification based on user interaction, classification based on user feedback, and classification based on time series.
[0086] In the classification based on purchase behavior, users can be classified based on two dimensions: purchase frequency and purchase amount.
[0087] When classified by purchase frequency, users can be divided into high-frequency users, medium-frequency users and low-frequency users. Among them, high-frequency users can refer to users who purchase at least once a week, medium-frequency users can refer to users who purchase 1-3 times a month, and low-frequency users can refer to users who purchase once a quarter or less.
[0088] When classified by purchase amount, users can be divided into high-consumption users, medium-consumption users and low-consumption users. Among them, high-consumption users can refer to users whose purchase amount exceeds 500 yuan each time, medium-consumption users can refer to users whose purchase amount is between 100 yuan and 500 yuan each time, and low-consumption users can refer to users whose purchase amount is less than 100 yuan each time.
[0089] In the classification based on user attributes, users can be classified based on two dimensions: user age and user geographic location.
[0090] When users are classified by age, they can be divided into adolescent users, young users, middle-aged users and elderly users. Among them, adolescent users refer to users aged 13 to 18, young users refer to users aged 19 to 35, middle-aged users refer to users aged 36 to 60, and elderly users refer to users over 60.
[0091] When classified by users' geographical location, users can be divided into first-tier city users, second-tier city users, third-tier city users and users in rural areas.
[0092] In the user interaction-based classification, users can be classified according to their activity level on the website. For example, users can be divided into active users, moderately active users, and inactive users. Among them, active users refer to users who log in to the website every day and browse multiple pages, moderately active users refer to users who log in to the website 2-3 times a week, and inactive users refer to users who log in to the website once a month or less.
[0093] In the classification based on user feedback, users can be classified according to their evaluation of the product. For example, users can be divided into high-satisfaction users, medium-satisfaction users, and low-satisfaction users. Among them, high-satisfaction users refer to users with an average rating of more than 4 stars, medium-satisfaction users refer to users with an average rating of 3-4 stars, and low-satisfaction users refer to users with an average rating of less than 3 stars.
[0094] In the time series-based classification, users are classified according to the periodicity of their product purchases. For example, users can be divided into regular purchasers, occasional purchasers, and very occasional purchasers. Regular purchasers refer to users who purchase on a fixed date each month, occasional purchasers refer to users who purchase once a quarter, and very occasional purchasers refer to users who purchase once a year or less.
[0095] It can be understood that the above description of different dimensions of user classification is only an example. In actual applications, the dimensions of user classification can be set according to different needs and are not specifically limited here.
[0096] After obtaining the classification results of users, the user classification module 130 can determine the push strategies for different categories of users based on the set push strategies and classification results. For example, membership plans can be launched for high-frequency users, and promotional information can be sent to low-frequency users. In this way, not only can user satisfaction be increased, but also the conversion rate and repurchase rate of users can be improved.
[0097] In this embodiment, the data collection module 110 collects marketing data from a real-time data source to achieve efficient collection of marketing data. After the marketing data is collected, the marketing data is preprocessed in parallel in an asynchronous manner, which can increase the processing speed of the marketing data. Next, the data analysis module 120 can obtain the preprocessed marketing data from the database, and identify the authenticity of the marketing data according to the preset recognition model analysis, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior according to the recognition result of the authenticity of the marketing data. When it is determined that the merchant corresponding to the marketing data has false marketing behavior, the merchant can be marked to reduce the frequency of recommendation for the merchant. At the same time, the user classification module 130 can classify users according to the marketing data obtained from the database using a preset classification model to obtain accurate user categories, so as to achieve accurate push for users of different categories according to the classification results and the set push strategy.
[0098] In one possible implementation, Figure 2 As shown, the data acquisition module 110 may include a real-time data acquisition unit 1101 and an asynchronous data processing unit 1102 .
[0099] Specifically, the real-time data collection unit 1101 is used to collect marketing data from the server corresponding to the real-time data source, and the real-time data source includes at least social media, e-commerce platforms and real-time log streams, so as to obtain marketing data reflecting the latest market dynamics and consumer behaviors. The asynchronous data processing unit 1102 is used to asynchronously pre-process the marketing data collected by the real-time data collection unit 1101, and store the pre-processed marketing data in the database, thereby improving the data processing speed by processing the marketing data in parallel.
[0100] Considering that the Kafka server, as a distributed stream processing platform, has the characteristics of high performance, persistence, distributed structure, publish-subscribe model, fault tolerance and horizontal scalability, in this embodiment, the server corresponding to the real-time data source can be a Kafka server. Specifically, the real-time data collection unit 1101 can subscribe to the marketing data of social media, the marketing data of the e-commerce platform and the marketing data of the implementation log stream from the Kafka server; loop through the messages received from the Kafka server, extract the marketing data from the messages; and send the extracted marketing data to the asynchronous data processing unit 1102 for processing.
[0101] Specifically, the real-time data collection unit 1101 can subscribe to the marketing data of the required real-time data source according to the address and port of the specified Kafka server, for example, the marketing data of social media, the marketing data of the e-commerce platform, and the marketing data implementing the log stream. The Kafka server will publish the marketing data about social media, the marketing data of the e-commerce platform, and the marketing data implementing the log stream, and the real-time data collection unit 1101 receives the subscribed message, which may include the marketing data of social media, the marketing data of the e-commerce platform, or the marketing data implementing the log stream. The real-time data collection unit 1101 extracts the marketing data from the message, and then sends the extracted marketing data to the asynchronous data processing unit 1102.
[0102] The asynchronous data processing unit 1102 receives the marketing data from the real-time data collection unit 1101, performs data cleaning and data format conversion on the marketing data through the asynchronous task function, and stores the processed marketing data into the database through the data saving function; among which, the data cleaning and data format conversion belong to preprocessing.
[0103] In this way, by subscribing to marketing data from various real-time data sources from the Kafka server, the real-time and diversity of the collected marketing data can be guaranteed, thereby achieving more accurate user classification and false behavior judgment. In addition, marketing data can be processed in parallel during the data collection process, improving data processing speed.
[0104] In one possible implementation, when the data analysis module 120 in the above embodiment analyzes and identifies the authenticity of the marketing data according to a preset recognition model, it can extract the text features and numerical features of the marketing data through the preset recognition model; and calculate the term frequency-inverse document frequency (TF-IDF) value and the sentiment score corresponding to the text features through the preset recognition model, calculate the text length of the marketing data according to the numerical features and extract the timestamp features, and encode the product categories and user levels in the marketing data to obtain category features; obtain a feature matrix based on the text features, numerical features, and category features; and obtain the recognition result of the authenticity of the marketing data based on the feature matrix, the TF-IDF value, and the sentiment score.
[0105] If the marketing data currently analyzed and identified is the user comment data of a certain e-commerce platform, after obtaining the user comment data, the data analysis module 120 extracts the text features and numerical features of the user comment data through the preset recognition model, and calculates the TF-IDF value corresponding to the text features to obtain the frequency of each word in the comment, and uses the sentiment analysis tool to calculate the sentiment score corresponding to the text features of the user comment data. At the same time, according to the numerical features of the user comment data, the text length of the user comment data, that is, the number of words or characters, is calculated, and the timestamp feature is extracted, and the timestamp feature can be further converted into the number of hours in a day, the day of the week, etc. In addition, the product category of the product involved in the user comment data and the user level of the user who issued the user comment data on the e-commerce platform are encoded to obtain the category feature. Afterwards, a feature matrix is obtained based on the text features, numerical features, and category features, so that the authenticity recognition result of the user comment data is obtained based on the feature matrix, TF-IDF value, and sentiment score.
[0106] In the technical solution of this embodiment, by extracting the features of the marketing data and calculating the TF-IDF value and sentiment score, the authenticity recognition result of the marketing data is obtained, so that the system can determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result, and then screen out merchants that have engaged in false marketing behavior.
[0107] In a possible implementation, the user classification module 130 obtains marketing data from a database, and classifies users according to a preset classification model to obtain classification results, including: extracting user purchase record feature data from the marketing data; the user purchase record feature data includes user tag features, purchase frequency features, purchase amount features, and geographic location features. The user purchase record feature data is standardized to obtain user data to be classified. The user data to be classified is clustered and classified using a preset classification model to obtain a classification result.
[0108] In actual applications, after obtaining marketing data from the database, the user classification module 130 may first clean and organize the obtained marketing data to remove abnormal values and missing values in the marketing data. Then, the user's purchase record feature data is extracted from the cleaned and organized marketing data, including user tag features (such as user account name or account ID), purchase frequency features, purchase amount features, and geographic location features.
[0109] After obtaining the user's purchase record feature data, the purchase record feature data is scaled to achieve standardization of the purchase record feature data and obtain the user data to be classified. The standardization of the user's purchase record feature data can be achieved through a formula, that is, through the formula The user purchase record feature data is standardized to obtain the user data to be classified. Among them, z represents the user data to be classified, x represents the user purchase record feature data, μ represents the mean of each feature in the user purchase record feature data, that is, the mean of the user mark feature, purchase frequency feature, purchase amount feature and geographical location feature, and σ represents the standard deviation of each feature in the user purchase record feature data, that is, the standard deviation of the user mark feature, purchase frequency feature, purchase amount feature and geographical location feature.
[0110] Afterwards, the user classification module 130 clusters and classifies the user data to be classified using a preset classification model. The preset classification model may be pre-trained based on a clustering algorithm or a classification algorithm. The clustering algorithm may be, for example, a K-means clustering algorithm or a hierarchical clustering algorithm, and the classification algorithm may be, for example, a decision tree algorithm or a random forest algorithm. After obtaining the classification result of the user, the personalized marketing strategy of the user may be determined based on the classification result and the set push strategy.
[0111] It can be understood that by standardizing the feature data of user purchase records, the numerical ranges and scales of different features are made more consistent, thereby preventing certain features from affecting the reliability of the preset classification model due to a large numerical range.
[0112] Taking the use of K-means clustering algorithm to classify users according to their purchase record feature data as an example, the user classification module 130 obtains marketing data, performs data preprocessing on the marketing data, i.e., data cleaning and sorting, and deletes abnormal values and missing values in the marketing data. After that, feature extraction is performed on the preprocessed marketing data to extract purchase record feature data in the marketing data, including user age features, purchase frequency features, purchase amount features, and geographic location features.
[0113] After that, calculate the mean of user age characteristics, purchase frequency characteristics, purchase amount characteristics and geographical location characteristics, as well as the standard deviation of user age characteristics, purchase frequency characteristics, purchase amount characteristics and geographical location characteristics, and then, according to the formula The user purchase record feature data is standardized to obtain the user data to be classified.
[0114] Assume that the characteristic values of the user's purchase record characteristic data are as follows:
[0115] Purchase frequency Purchase amount User age Location 1 50 25 0 2 100 30 1 3 150 35 2 4 200 40 3
[0116] After normalization, the resulting eigenvalues may appear as shown in the following table:
[0117] Purchase frequency Purchase amount User age Location -0.5 -1.0 -1.0 -1.0 0.0 0.0 0.0 0.0 0.5 1.0 1.0 1.0 1.0 2.0 2.0 2.0
[0118] In this example, each feature value is scaled to have a mean of 0 and a standard deviation of 1, making the range and scale of values across features more consistent.
[0119] The data to be classified is input into the preset classification model, and the features of each cluster center are output as the classification result. One cluster center represents a category of users. Thus, multi-dimensional classification of users is achieved.
[0120] In the above scheme, the user classification module 130 can extract user purchase record feature data from the marketing data, and obtain classification results about the user based on the user purchase record feature data, thereby realizing multi-dimensional classification of users, which helps to understand the user group more carefully and thus achieve more accurate push.
[0121] Figure 3 FIG. 2 shows a schematic diagram of the structure of another marketing data analysis system according to an embodiment of the present invention. Figure 3 As shown, on the basis of the above embodiment, in order to protect the marketing data analysis system 100 from the risks of network attacks and data leakage, the marketing data analysis system 100 provided in this embodiment can further include a data security module 140, which is used to issue an alarm message when a network attack behavior is identified, and record a network attack log to facilitate security audits.
[0122] In actual applications, the data security module 140 can achieve comprehensive protection of the marketing data analysis system 100 through encryption technology, intrusion detection, firewalls and access control, security auditing and monitoring, backup and recovery strategies, and user education and training.
[0123] For example, encryption technology may include data encryption and transmission encryption. Data encryption may use encryption algorithms to encrypt sensitive data to ensure that the data cannot be easily decrypted even if it is intercepted. Transmission encryption may use HTTPs protocol and other methods to protect the security of marketing data during transmission.
[0124] Intrusion detection can monitor network traffic in real time and identify abnormal traffic patterns. It can also analyze system logs to detect unauthorized access attempts and other potential security threats. It can also perform behavioral analysis to detect and identify abnormal behavior patterns by learning normal behavior patterns to provide prompts for potential attacks.
[0125] Firewalls and access control can configure the system with a network firewall to block traffic from known malicious IP addresses, protect web applications from attacks through an application firewall, and limit access to sensitive data resources through authentication and authorization.
[0126] Security auditing and monitoring can regularly conduct security audits on the system, check the security of the system, and fix any discovered security vulnerabilities. At the same time, through real-time monitoring of the system, an immediate response can be issued when a security incident occurs.
[0127] The backup and recovery strategy can achieve regular backup of marketing data and ensure the safe storage of backup data. At the same time, by formulating a disaster recovery plan, it can be quickly restored in the event of marketing data loss or system failure.
[0128] User education and training can be achieved by regularly conducting cybersecurity awareness training for employees to avoid security risks caused by mistakes and providing employees with security operation guides to guide them in properly handling sensitive information.
[0129] In this way, data security protection is achieved through the data security module 140, and an alarm message can be issued when a network attack behavior is identified, and a network attack log can be recorded, such as the time, type, source and other information of the network attack, so that relevant personnel can take necessary measures in a timely manner to find potential security loopholes.
[0130] In one implementation, the data security module 140 can start a monitoring thread to monitor the data flow of the network port corresponding to the monitoring address and port according to the set monitoring address and port, and obtain monitoring data. When it is determined according to the monitoring data that a preset network attack behavior occurs, an alarm message is issued, a network attack log is recorded, and an interrupt signal is issued to disconnect the network connection.
[0131] Among them, the alarm information may include sending information in the form of emails, text messages, etc. to alert relevant staff members of the occurrence of cyber attacks.
[0132] Optionally, when a preset network attack behavior is identified, a network attack log is recorded, that is, the time, type, source and other information of the attack are recorded, and relevant logs and data packets are retained. At the same time, the network connection is disconnected to isolate the affected system from the network to prevent the attack from spreading, and ports and services that may lead to further attacks can be closed.
[0133] Furthermore, you can also try to trace the source of the attack to gain a deeper understanding of the means used by the attacker and analyze the vulnerabilities in the system that may have led to the successful attack. After that, install security updates and patches on the affected systems and change the configuration to close unnecessary services and ports.
[0134] In the above scheme, the data security module 140 starts a monitoring thread to listen to the set listening address and port to obtain the listening data, so that it can determine whether a preset network attack behavior is currently occurring based on the listening data, and when it is determined that a preset network attack behavior has occurred, the network connection is promptly disconnected through an interrupt signal to prevent the attack from spreading.
[0135] The technical solution of this embodiment collects marketing data from real-time data sources through the data collection module 110, realizes efficient collection of marketing data, and after collecting the marketing data, pre-processes the marketing data in parallel in an asynchronous manner, which can increase the processing speed of the marketing data. Then, the data analysis module 120 can obtain the pre-processed marketing data from the database, and identify the authenticity of the marketing data according to the preset recognition model analysis, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior according to the recognition result of the authenticity of the marketing data, and when it is determined that the merchant corresponding to the marketing data has false marketing behavior, the merchant can be marked to reduce the frequency of recommendation for the merchant. At the same time, the user classification module 130 can classify users according to the marketing data obtained from the database using the preset classification model to obtain accurate user categories, so as to realize accurate push for users of different categories according to the classification results and the set push strategy.
[0136] Figure 4 FIG. 1 is a flow chart of a marketing data analysis method provided by an embodiment of the present invention. Figure 4 As shown, the marketing data analysis method provided in this embodiment can be applied to the marketing data analysis system 100 as described in any of the above embodiments. Specifically, the marketing data analysis method of this embodiment may include the following steps:
[0137] 4100. The data collection module 110 in the marketing data analysis system 100 collects marketing data from real-time data sources.
[0138] Among them, the data collection module 110 can collect marketing data from the server corresponding to the real-time data source, and the real-time data source includes at least social media, e-commerce platforms and real-time log streams, so as to obtain marketing data reflecting the latest market dynamics and consumer behaviors.
[0139] 4200. The data collection module 110 pre-processes the marketing data and stores the pre-processed marketing data in a database.
[0140] The preprocessing of marketing data may include data cleaning and data format conversion. Data cleaning is to remove noise data from marketing data, and data format conversion can standardize and normalize marketing data to improve the accuracy of subsequent processing.
[0141] The data collection module 110 can asynchronously pre-process the collected marketing data and store the pre-processed marketing data in a database, thereby improving the data processing speed by processing the marketing data in parallel.
[0142] In one implementation, the data collection module 110 can subscribe to the marketing data of the required real-time data source, such as marketing data of social media, marketing data of e-commerce platforms, and marketing data implementing log streams, according to the address and port of the specified Kafka server. The Kafka server will publish marketing data about social media, marketing data of e-commerce platforms, and marketing data implementing log streams, and the data collection module 110 receives subscribed messages, which may include marketing data of social media, marketing data of e-commerce platforms, or marketing data implementing log streams. The data collection module 110 extracts the marketing data from the message, performs data cleaning and data format conversion on the marketing data through an asynchronous task function, and stores the processed marketing data in the database through a data preservation function.
[0143] In this way, by subscribing to marketing data from various real-time data sources from the Kafka server, the real-time and diversity of the collected marketing data can be guaranteed, thereby achieving more accurate user classification and false behavior judgment. In addition, marketing data can be processed in parallel during the data collection process, improving data processing speed.
[0144] 4300. The data analysis module 120 in the marketing data analysis system 100 obtains marketing data from the database, and analyzes and identifies the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior based on the recognition result of the authenticity of the marketing data.
[0145] When the data analysis module 120 analyzes and identifies the authenticity of the marketing data according to the preset recognition model, it can extract the text features and numerical features of the marketing data through the preset recognition model; and calculate the term frequency-inverse document frequency (TF-IDF) value and the sentiment score corresponding to the text features through the preset recognition model, calculate the text length of the marketing data according to the numerical features and extract the timestamp features, and encode the product categories and user levels in the marketing data to obtain category features; obtain a feature matrix according to the text features, numerical features, and category features; and obtain the recognition result of the authenticity of the marketing data based on the feature matrix, TF-IDF value, and sentiment score.
[0146] 4400. The user classification module 130 in the marketing data analysis system 100 obtains marketing data from the database, classifies users according to a preset classification model to obtain classification results, and determines push strategies for different categories of users based on the set push strategies and classification results.
[0147] Specifically, the user classification module 130 can extract user purchase record feature data from the marketing data; the user purchase record feature data includes user tag features, purchase frequency features, purchase amount features, and geographic location features. The user purchase record feature data is standardized to obtain user data to be classified. The user data to be classified is clustered and classified using a preset classification model to obtain a classification result.
[0148] In actual applications, after obtaining marketing data from the database, the user classification module 130 may first clean and organize the obtained marketing data to remove abnormal values and missing values in the marketing data. Then, the user's purchase record feature data is extracted from the cleaned and organized marketing data, including user tag features (such as user account name or account ID), purchase frequency features, purchase amount features, and geographic location features.
[0149] After obtaining the user's purchase record feature data, the purchase record feature data is scaled to achieve standardization of the purchase record feature data and obtain the user data to be classified. The standardization of the user's purchase record feature data can be achieved through a formula, that is, through the formula The user purchase record feature data is standardized to obtain the user data to be classified. Wherein, z represents the user data to be classified, x represents the user purchase record feature data, μ represents the mean of each feature in the user purchase record feature data, that is, the mean of the user mark feature, purchase frequency feature, purchase amount feature and geographic location feature, and σ represents the standard deviation of each feature in the user purchase record feature data, that is, the standard deviation of the user mark feature, purchase frequency feature, purchase amount feature and geographic location feature. In this way, the numerical range and scale between different features can be made more consistent, thereby avoiding that some features affect the reliability of the preset classification model due to the large numerical range.
[0150] Afterwards, the user classification module 130 clusters and classifies the user data to be classified using a preset classification model. The preset classification model may be pre-trained based on a clustering algorithm or a classification algorithm. The clustering algorithm may be, for example, a K-means clustering algorithm or a hierarchical clustering algorithm, and the classification algorithm may be, for example, a decision tree algorithm or a random forest algorithm. After obtaining the classification result of the user, the personalized marketing strategy of the user may be determined based on the classification result and the set push strategy.
[0151] In the technical solution of this embodiment, the data collection module 110 is used to collect marketing data from a real-time data source, pre-process the marketing data, and store the pre-processed marketing data in a database. The data analysis module 120 is used to obtain marketing data from the database, and analyze and identify the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has false marketing behavior based on the recognition result. The user classification module 130 is used to obtain marketing data from the database, classify users according to a preset classification model to obtain classification results, and determine the push strategy for different categories of users based on the set push strategy and classification results. This allows faster and more efficient processing of massive marketing data and accurate classification of users.
[0152] Furthermore, the marketing data analysis method provided by the embodiment of the present invention may also include: the data security module 140 of the marketing data analysis system 100 issues an alarm message when a network attack behavior is identified, and records a network attack log to facilitate security auditing.
[0153] By issuing an alarm message when a network attack behavior is identified, the marketing data analysis system 100 can be protected from the risks of network attacks and data leakage, thereby achieving comprehensive protection of the marketing data analysis system 100.
[0154] In a specific implementation, a monitoring thread can be started according to the set listening address and port to monitor the data flow of the network port corresponding to the listening address and port to obtain the monitoring data. When it is determined according to the monitoring data that a preset network attack behavior occurs, an alarm message is issued, a network attack log is recorded, and an interrupt signal is issued to disconnect the network connection.
[0155] Among them, the alarm information may include sending information in the form of emails, text messages, etc. to alert relevant staff members of the occurrence of cyber attacks.
[0156] Optionally, when a preset network attack behavior is identified, a network attack log is recorded, that is, the time, type, source and other information of the attack are recorded, and relevant logs and data packets are retained. At the same time, the network connection is disconnected to isolate the affected system from the network to prevent the attack from spreading, and ports and services that may lead to further attacks can be closed.
[0157] Furthermore, you can also try to trace the source of the attack to gain a deeper understanding of the means used by the attacker and analyze the vulnerabilities in the system that may have led to the successful attack. After that, install security updates and patches on the affected systems and change the configuration to close unnecessary services and ports.
[0158] In the technical solution of this embodiment, the data security module 140 starts a monitoring thread to listen to the set listening address and port to obtain the listening data, so that it can determine whether a preset network attack behavior is currently occurring based on the listening data, and when it is determined that a preset network attack behavior has occurred, the network connection is promptly disconnected through an interrupt signal to prevent the attack from spreading.
[0159] Figure 5 FIG. 1 is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present invention. Figure 5 As shown, an electronic device 500 provided in this embodiment includes: a processor 501 and a memory 502; wherein:
[0160] The memory 502 is used to store computer programs, and the memory may also be a flash memory.
[0161] The processor 501 is used to execute the execution instructions stored in the memory to implement each step in the above method. For details, please refer to the relevant description in the above method embodiment.
[0162] Optionally, the memory 502 may be independent or integrated with the processor 501 .
[0163] When the memory 502 is a device independent of the processor 501, the electronic device 500 may further include:
[0164] The bus 503 is used to connect the memory 502 and the processor 501 .
[0165] This embodiment also provides a computer-readable storage medium, in which a computer program is stored. When at least one processor of an electronic device executes the computer program, the electronic device executes the methods provided in the above-mentioned various embodiments.
[0166] This embodiment also provides a program product, which includes a computer program stored in a readable storage medium. At least one processor of the electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device implements the methods provided in the above various embodiments.
[0167] Those skilled in the art will readily appreciate other implementations of the embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The embodiments of the present invention are intended to cover any variations, uses, or adaptations of the embodiments of the present invention that follow the general principles of the embodiments of the present invention and include common knowledge or customary technical means in the art that are not disclosed in the embodiments of the present invention. The specification and embodiments are to be regarded as exemplary only, and the true scope and spirit of the embodiments of the present invention are indicated by the claims.
[0168] It should be understood that the embodiments of the present invention are not limited to the precise structures described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the embodiments of the present invention is limited only by the appended claims.
Claims
1. A marketing data analysis system, characterized in that: include: A data collection module is used to collect marketing data from real-time data sources, pre-process the marketing data, and store the pre-processed marketing data in a database; A data analysis module, used to obtain the marketing data from the database, and analyze and identify the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing behavior according to the recognition result of the authenticity of the marketing data; The user classification module is used to obtain the marketing data from the database, classify users according to a preset classification model to obtain classification results, and determine the push strategies for different categories of users based on the set push strategies and the classification results.
2. The system according to claim 1, characterized in that The data acquisition module comprises: A real-time data collection unit, configured to collect the marketing data from a server corresponding to the real-time data source, wherein the real-time data source includes at least social media, an e-commerce platform, and a real-time log stream; The asynchronous data processing unit is used to asynchronously pre-process the marketing data collected by the real-time data collection unit and store the pre-processed marketing data in the database.
3. The system according to claim 2, characterized in that The real-time data collection unit is specifically used to: subscribe to the marketing data of the social media, the marketing data of the e-commerce platform and the marketing data of the implementation log stream from the Kafka server; loop through the messages received from the Kafka server and extract the marketing data from the messages; and send the extracted marketing data to the asynchronous data processing unit for processing.
4. The system according to claim 2, characterized in that The asynchronous data processing unit is specifically used for: receiving the marketing data from the real-time data acquisition unit, performing data cleaning processing and data format conversion processing on the marketing data through an asynchronous task function, and storing the processed marketing data to the database through a data saving function; wherein the data cleaning processing and the data format conversion processing belong to the preprocessing.
5. The system according to claim 1, characterized in that The data analysis module analyzes and identifies the authenticity of the marketing data according to a preset recognition model, including: Extracting text features and numerical features of the marketing data through the preset recognition model; Calculating the term frequency-inverse document frequency TF-IDF value and sentiment score corresponding to the text feature through the preset recognition model, calculating the text length of the marketing data according to the numerical feature and extracting the timestamp feature, and encoding the product category and user level in the marketing data to obtain the category feature; Obtain a feature matrix according to the text features, the numerical features, and the category features; Based on the feature matrix, the TF-IDF value and the sentiment score, an identification result of the authenticity of the marketing data is obtained.
6. The system according to claim 1, characterized in that The user classification module obtains the marketing data from the database and classifies the users according to the preset classification model to obtain classification results including: Extracting user purchase record feature data from the marketing data; the user purchase record feature data includes user tag features, purchase frequency features, purchase amount features and geographic location features; Standardizing the user purchase record feature data to obtain user data to be classified; The user data to be classified is clustered and classified by using the preset classification model to obtain the classification result.
7. The system according to claim 6, characterized in that The user classification module performs standardization processing on the user purchase record feature data to obtain user data to be classified, including: By formula Standardizing the user purchase record feature data to obtain the user data to be classified; Wherein, z represents the user data to be classified, x represents the user purchase record feature data, μ represents the mean of each feature in the user purchase record feature data, and σ represents the standard deviation of each feature in the user purchase record feature data.
8. The system according to claim 1, characterized in that The system also includes: a data security module, which is used to issue an alarm message when a network attack behavior is identified, and record a network attack log to facilitate security auditing.
9. The system according to claim 8, characterized in that The data security module is specifically used to: start a monitoring thread to monitor the data of the network port corresponding to the monitoring address and port according to the set monitoring address and port, and obtain the monitoring data; When it is determined according to the monitoring data that the preset network attack behavior occurs, the alarm information is issued, the network attack log is recorded, and an interrupt signal is issued to disconnect the network connection.
10. A marketing data analysis method, applied to the marketing data analysis system according to any one of claims 1 to 9, characterized in that: The method comprises: The data collection module in the marketing data analysis system collects marketing data from real-time data sources; The data collection module pre-processes the marketing data and stores the pre-processed marketing data in a database; The data analysis module in the marketing data analysis system obtains the marketing data from the database, and analyzes and identifies the authenticity of the marketing data according to a preset recognition model, so as to determine whether the merchant corresponding to the marketing data has engaged in false marketing activities according to the recognition result of the authenticity of the marketing data; The user classification module in the marketing data analysis system obtains the marketing data from the database, classifies users according to a preset classification model to obtain classification results, and determines push strategies for different categories of users based on set push strategies and the classification results.
Citation Information
Patent Citations
Enterprise false publicity risk identification method and device, and storage medium
CN113505221A
Marketing content inspecting apparatus and method
CN113657918A
Ship abnormal behavior identification method and device, terminal and storage medium
CN114742128A
Automatic software test management system and method based on artificial intelligence
CN114819985A
One-stop datamation intelligent marketing system
CN118096204A
