Information publishing system based on big data and two-dimensional code hierarchical management

Abstract

The invention relates to the field of information release, and discloses an information release system based on big data and two-dimensional code hierarchical management, which comprises a multi-level two-dimensional code generation engine, a user behavior data acquisition module and an intelligent recommendation module, the multi-level two-dimensional code generation engine is used for generating a public code and a dynamic level private code based on the public code according to client registration information; the user behavior data acquisition module is used for recording a user code scanning path in real time, including a device type, a code scanning hierarchy and a page stay duration, and constructing a user behavior tag; and the intelligent recommendation module is used for carrying out single-category recommendation on anonymous users and carrying out multi-category recommendation on real-name users by adopting a dynamic recommendation strategy based on the user portraits and the user behavior tags, and displaying recommendation contents at a PC terminal and a mobile terminal.

Description

Technical Field

[0001] The present invention relates to the field of information publishing, and in particular to an information publishing system based on big data and hierarchical management of QR codes. Background Art

[0002] With the rapid development of information technology and the popularization of mobile Internet, the way of information release and acquisition has undergone profound changes. Traditional information release systems often adopt a single, static approach, which is difficult to meet the growing personalized and diversified needs of users. At the same time, with the rise of big data technology, how to effectively use massive user behavior data to achieve accurate information push and personalized services has become a key issue that needs to be solved in the current information release field.

[0003] In the prior art, QR codes, as a convenient information carrier and interactive portal, have been widely used in information release, product promotion, user authentication, etc. However, traditional QR code systems can usually only achieve single-level information links and lack the ability of hierarchical management and dynamic recommendation. This results in users being unable to obtain more accurate and personalized information recommendations based on their interests and needs during the QR code scanning process. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an information publishing system based on big data and hierarchical management of QR codes, including a multi-level QR code generation engine, a user behavior data collection module and an intelligent recommendation module;

[0005] The multi-level QR code generation engine is used to generate a public code and a dynamic hierarchical private code based on the public code according to the customer registration information;

[0006] User behavior data collection module, which is used to record the user's scanning path in real time, including device type, scanning level and page dwell time, and build user behavior tags;

[0007] The intelligent recommendation module adopts a dynamic recommendation strategy based on user portraits and user behavior tags, makes single-category recommendations to anonymous users and multi-category recommendations to real-name users, and displays recommended content on PC and mobile terminals.

[0008] Furthermore, the generation of a public code and a dynamic hierarchical private code based on the public code according to the customer registration information includes:

[0009] The public code is a two-dimensional code generated according to user information, and the dynamic hierarchical private code includes multiple layers of subcodes, each layer of subcodes is generated based on the upper layer of subcodes, and the subcode is a two-dimensional code formed by the public code plus a classification identifier plus a hash encryption sequence.

[0010] Further, the user behavior tags include the interest intensity grading based on the page stay duration and the category preferences associated with the scanned private code levels.

[0011] Further, the interest intensity grading based on the page stay duration includes the scanning code path analysis, constructing the user behavior chain, and calculating the interest weights;

[0012] Among them, the scanning code path analysis is to obtain the scanning code path according to the scanning order within the same session; based on the information of each sub-code and the stay duration on the scanning code path, construct the user behavior chain, and obtain the category preferences and category interest weights according to the user behavior chain; the category interest weight is the proportion of the total stay duration of each sub-code under the same public code in the stay duration, and the interest weight of the category associated with the public code is obtained.

[0013] Further, the category preferences are obtained according to the user behavior chain, where the category preferences include single-category preferences and multi-category preferences; the single-category preference is that within the same session, the number of sub-codes scanned under the same public code is not less than the set number and the total stay duration is greater than the set duration;

[0014] The multi-category preference is that the number of times of scanning sub-codes across public codes is not less than the set number and the proportion of the stay duration under each public code is not less than the set proportion.

[0015] Further, based on the user portrait and user behavior tags, a dynamic recommendation strategy is adopted to perform single-category recommendations for anonymous users and multi-category recommendations for real-name users, and display the recommended content on the PC side and the mobile side, including:

[0016] The recommendation engine generates a main container, a sub-container, and a dynamic sub-container. The main container binds the public code to store the customer basic portrait and the global recommendation strategy. The sub-container associates with the sub-code classification to execute the category-level recommendation rules. The dynamic sub-container performs category recommendations according to the user behavior tags.

[0017] The beneficial effects of the present invention are as follows: Through the multi-level two-dimensional code generation engine and the user behavior data acquisition module, the present invention realizes the in-depth tracking and analysis of user behavior. Based on this, the intelligent recommendation module can provide more accurate and personalized recommended content for users, significantly improving the accuracy and conversion rate of marketing. By collecting and analyzing user behavior data in real time, the present invention can dynamically adjust the recommendation strategy to ensure that the recommended content always meets the current interests and needs of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the principle of an information publishing system based on big data and two-dimensional code hierarchical management;

[0019] Figure 2Schematic diagram of the recommendation method process based on big data and two-dimensional code hierarchical management;

[0020] Figure 3 Schematic diagram of the implementation of the information release system based on big data and two-dimensional code hierarchical management. Detailed implementation manners

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0022] The features and performance of the present invention will be further described in detail below with reference to the embodiments.

[0023] As Figure 1 shown, the information release system based on big data and two-dimensional code hierarchical management includes a multi-level two-dimensional code generation engine, a user behavior data collection module, and an intelligent recommendation module;

[0024] The multi-level two-dimensional code generation engine is used to generate a public code and dynamic hierarchical private codes based on the public code according to the customer registration information;

[0025] The user behavior data collection module is used to record the user's two-dimensional code scanning path in real time, including the device type, the scanning level, and the page stay duration, and construct user behavior tags;

[0026] The intelligent recommendation module, based on the user portrait and user behavior tags, adopts a dynamic recommendation strategy to perform single-category recommendations for anonymous users and multi-category recommendations for real-name users, and displays the recommended content on the PC side and the mobile side.

[0027] The generation of the public code and the dynamic hierarchical private codes based on the public code according to the customer registration information includes:

[0028] The public code is a two-dimensional code generated according to the user information. The dynamic hierarchical private codes include multiple layers of sub-codes. Each layer of sub-code is generated based on the upper-layer sub-code. The sub-code is a two-dimensional code formed by adding a classification identifier and a hash encryption sequence to the public code.

[0029] The user behavior tags include the interest intensity classification based on the page stay duration and the category preference associated with the scanning level of the private code.

[0030] The interest intensity classification based on the page stay duration includes two-dimensional code scanning path analysis, constructing a user behavior chain, and interest weight calculation;

[0031] The parsing of the code scanning path is to obtain the code scanning path according to the sequence of code scanning within the same session; based on the information of each sub-code on the code scanning path and the residence duration, construct a user behavior chain, and obtain the category preference and category interest weight according to the user behavior chain; the category interest weight is the proportion of the total residence duration of each sub-code under the same public code in the residence duration, to obtain the interest weight of the category associated with the public code.

[0032] The obtaining of the category preference according to the user behavior chain, where the category preference includes single-category preference and multi-category preference; the single-category preference is that within the same session, scan no less than a set number of sub-codes under the same public code and the total residence duration is greater than the set duration;

[0033] The multi-category preference is that the number of times of scanning sub-codes across public codes is not less than the set number of times and the proportion of the residence duration under each public code is not less than the set proportion.

[0034] Based on the user portrait and user behavior tags, adopt a dynamic recommendation strategy to perform single-category recommendations for anonymous users and multi-category recommendations for real-name users, and display the recommended content on the PC side and the mobile side, including:

[0035] The recommendation engine generates a main container, a sub-container, and a dynamic sub-container. The main container binds the public code to store the customer basic portrait and the global recommendation strategy. The sub-container associates the sub-code classification to execute the category-level recommendation rules. The dynamic sub-container performs category recommendations according to the user behavior tags.

[0036] As Figure 2 shown, the recommendation method based on big data and two-dimensional code hierarchical management is applied to the information publishing system based on big data and two-dimensional code hierarchical management, including generating multi-level two-dimensional codes, collecting user behavior data, constructing user portraits, executing dynamic recommendation strategies, displaying recommended content, real-time feedback correction, and containerized operation management.

[0037] Specifically, the present invention proposes an information publishing system and method based on big data and two-dimensional code hierarchical management, aiming to achieve in-depth insight into user behavior and accurate recommendation through the collaborative work of a multi-level two-dimensional code generation engine, a user behavior data collection module, and an intelligent recommendation module.

[0038] Multi-level two-dimensional code generation engine

[0039] The multi-level QR code generation engine of the present invention is responsible for generating public codes and dynamic hierarchical private codes based on the public codes according to the customer registration information. The public code is a unique identification QR code generated according to the basic information of the user (or customer, such as a merchant), serving as the basic entry point for the entire marketing system. The dynamic hierarchical private code is a multi-level sub-code system generated by adding classification identifiers and hash encryption sequences on the basis of the public code. Each layer of sub-codes is generated based on its upper-layer sub-code, forming a tree-like or chain-like structure, which is convenient for tracking the user's behavior path between different levels.

[0040] Public code generation: According to the registration information provided by the customer (such as enterprise name, account ID, etc.), a unique public code is generated through specific coding rules. The public code serves as the identification of the customer's identity, associating with the customer's basic portrait and global recommendation strategy.

[0041] Dynamic hierarchical private code generation: On the basis of the public code, according to the needs of marketing activities, different classification identifiers (such as product categories, activity themes, etc.) are designed, and combined with the hash encryption algorithm to generate multi-level sub-codes. The sub-codes not only inherit the basic information of the public code, but also carry more fine-grained classification information, which is convenient for subsequent user behavior analysis and precise recommendation.

[0042] User behavior data collection module

[0043] The user behavior data collection module is responsible for real-time recording of the user's QR code scanning path, including device type, scanning level, and page stay duration, and constructing user behavior tags based on these data.

[0044] Scanning path recording: Through buried point technology, the scanning order of the user within the same session is tracked to form a scanning path. The scanning path reflects the user's jumping behavior between different hierarchical private codes and is an important basis for analyzing the user's interest preferences.

[0045] User behavior chain construction: Based on the information of each sub-code on the scanning path and the user's stay duration on the page, a user behavior chain is constructed. The user behavior chain not only contains the sequential information of the user's access, but also incorporates the time dimension, making the description of the user's behavior more three-dimensional.

[0046] User behavior tag construction: According to the user behavior chain, the interest intensity grading based on the page stay duration and the category preference associated with the scanned private code level are extracted. The interest intensity grading is evaluated by calculating the proportion of the stay duration of the user on a specific category, while the category preference is determined according to the scanning frequency and stay duration of the user between different categories.

[0047] Intelligent recommendation module

[0048] Based on user profiles and user behavior tags, the intelligent recommendation module adopts a dynamic recommendation strategy to provide differential recommendations for anonymous users and real-name users, and displays the recommended content on the PC side and the mobile side.

[0049] User profile construction: Integrate the user's basic information (such as age, gender, geographical location, etc.), historical behavior data (such as purchase records, browsing history, etc.) and real-time behavior tags to construct a comprehensive user profile. The user profile is the basis for formulating personalized recommendation strategies.

[0050] Dynamic recommendation strategy: For anonymous users, due to the lack of historical behavior data, a single-category recommendation strategy is mainly adopted, that is, according to the user's current browsing category preferences, relevant products or activities are recommended. For real-name users, a multi-category recommendation strategy is adopted by combining their historical behavior data and real-time behavior tags to provide more personalized and diverse recommended content.

[0051] Display of recommended content: According to the user's device type (PC side or mobile side), optimize the display method and layout of the recommended content to ensure that the recommended information can be presented in the most attractive way to users.

[0052] Generate a public code based on the customer registration information and a dynamic hierarchical private code based on the public code

[0053] Process of generating the public code: When the customer registers, the system extracts their key information (such as enterprise name, account ID), and generates a unique public code through a hash function. The public code is stored in the database and associated with the customer's basic profile and global recommendation strategy.

[0054] Process of generating the dynamic hierarchical private code: Based on the public code, different classification identifiers (such as "electronic products", "clothing and shoes", etc.) are designed according to the needs of marketing activities. For each classification, the first-level sub-code is generated by adding the classification identifier and the hash encryption sequence. Subsequently, based on the first-level sub-code, the second-level, third-level, etc. sub-codes can be further generated to form a multi-level private code system.

[0055] Collection of user behavior data and construction of tags

[0056] Analysis of the QR code scanning path: When the user scans a QR code using a mobile device, the system records the scanning time, device type, and scanning level. By tracking the QR code scanning order within the same session, the user's QR code scanning path is obtained.

[0057] Construction of the user behavior chain: Based on the QR code scanning path, the system constructs the user behavior chain. For example, the user first scans the first-level sub-code of the "electronic products" classification, then browses the relevant page and stays for 5 minutes; then scans the first-level sub-code of the "clothing and shoes" classification and stays for 3 minutes. These behaviors are recorded to form the user behavior chain.

[0058] Interest intensity grading and category preference calculation: The system calculates the interest weight based on the user's stay duration on different category pages. For example, if the total stay duration of a user on the "electronic products" category page is 10 minutes and on the "clothing and shoes" category page is 5 minutes, the interest weight of the "electronic products" category is higher than that of the "clothing and shoes" category. At the same time, the system determines the user's category preference according to the level and frequency of the user scanning the private code.

[0059] Intelligent recommendation strategy and execution

[0060] Recommendation engine design: The recommendation engine consists of a main container, sub - containers, and dynamic sub - containers. The main container binds the public code and stores the customer's basic profile and global recommendation strategies; the sub - containers are associated with sub - code classifications and execute category - level recommendation rules; the dynamic sub - containers adjust and optimize the recommended content in real - time according to the user's behavior tags.

[0061] Implementation of dynamic recommendation strategy: For anonymous users, the system mainly recommends relevant products or activities based on their current browsing category preferences. For example, if a user is browsing a page in the "electronic products" category, the system recommends popular products or promotional activities in this category. For authenticated users, the system combines their historical purchase records, browsing history, and real - time behavior tags to provide more personalized recommended content. For example, if a user has historically purchased "electronic products" and currently shows strong interest in the "smart home" sub - category, the system recommends relevant smart home products.

[0062] Display and optimization of recommended content: The system optimizes the display method and layout of recommended content according to the user's device type (PC or mobile). On the PC side, methods such as sidebars, pop - ups, or embedded recommendations can be used; on the mobile side, methods such as sliding cards, bottom navigation bars, or floating buttons can be used. At the same time, the system continuously optimizes the recommendation algorithm and display method through methods such as A / B testing to improve the recommendation effect and user satisfaction.

[0063] As Figure 3 shown, the system generates a unique QR code for the user. For the same real - name authenticated customer, there is only one such QR code throughout their life, which is called the public code (main code). In addition, the system can generate private codes at different levels according to the needs of the same customer: named sub - codes, second - level sub - codes, third - level codes... etc.

[0064] The private codes of the same customer are generated by this system on a public code through the permutation and combination of Chinese characters + Arabic numerals + English letters, and theoretically, an infinite number of QR code private codes can be generated.

[0065] 3. Information aggregation

[0066] The customer uploads the prepared information to the system database, including but not limited to the company profile, corporate culture, brand profile, product description, product store, and web links that the customer needs to release externally, etc.

[0067] This system actively responds to the policies and requirements of regulatory agencies, adopts technologies and solutions such as big data and artificial intelligence, establishes an intelligent supervision system, and monitors and analyzes the information uploaded by customers in real time to ensure the legality and standardization of information dissemination activities.

[0068] 4. Information Processing

[0069] The system classifies the customer information that complies with laws and regulations, and assigns a QR code to each independent piece of information (including but not limited to company profile, corporate culture, brand profile, product description, product store, web links, etc.). The public code and private code relationship of the QR codes of the same customer is as follows:

[0070] A. 1 public code (main code): The only QR code for the customer throughout their lifetime.

[0071] B. 10 private codes (sub-codes): According to the information categories, they are divided into 10 major categories such as enterprise information, brand information, product information, activity information, industry news, etc. The customer can also customize the information classification.

[0072] C. 100 private codes (secondary sub-codes): 100 secondary sub-codes are assigned under each sub-code (i.e., classified information), and each independent piece of information obtains one secondary sub-code.

[0073] D. 1000 private codes (tertiary codes): 1000 tertiary codes are assigned under each secondary sub-code, and each independent piece of information obtains one tertiary code.

[0074] In this system, the total number of QR codes that a customer can use is:

[0075] 1 + 10 + 10×100 + 10×100×1000 = 1001011

[0076] One million QR codes can represent one million pieces of information, which is sufficient to meet the customer's information release needs.

[0077] 5. QR Code Application

[0078] The customer can publish the public code and private code of the QR codes generated by this system online and offline, and users can scan the QR codes to understand the relevant information.

[0079] 6. QR Code Application Instructions

[0080] The application of QR codes is divided into two situations:

[0081] 1) The system APP is not installed

[0082] Users can scan the public code and private code generated by this system with any third-party software with scanning function, and jump to the customer's information main page (home page). Users can search on the home page, or find the information they need to know according to the page menu.

[0083] Non-real-name registered users only have the right to browse all information publicly released by this system.

[0084] 2) The system APP has been installed

[0085] Users download and install the system APP (without real-name registration), use the APP to scan the public code, and jump to the customer information main page (home page). Scan the private code to jump to the corresponding independent information page. Users only have the right to browse all information publicly released by this system.

[0086] After users complete real-name registration on the APP, the membership function will be automatically activated. They can provide feedback and suggestions on the information posted by customers, or purchase customers' products at member prices.

[0087] Implementation Case 1: Application of Food Manufacturer Information Release System

[0088] A food manufacturer produces hundreds of varieties of products. By using this system, it can better realize the corporate culture, brand introduction and product information release, improve marketing efficiency and enhance user experience.

[0089] Multi-level QR code generation

[0090] Public code generation: Food manufacturers provide basic information such as company name and account ID to complete registration in this system. The system generates a unique public code through a hash function, and scanning the public code directly leads to the food manufacturer's information release homepage. The homepage can publish company profiles, corporate culture, brand profiles, promotional activities, etc., or other classified information that customers need to publish.

[0091] Private code generation: According to customer needs, design a category identifier under each category information, such as "cake", "bread", "biscuits", etc. For each category, generate the first-level subcode, i.e., the subcode, by adding the category identifier and the hash encryption sequence. For example, generate a subcode for the "cake" category, and on this basis, it can be subdivided into "birthday cake", "birthday cake", "children's cake", etc. according to the application scenario, and generate the second-level subcode, i.e., the secondary subcode. Under "birthday cake", it can be divided into "fruit birthday cake", "chocolate birthday cake", "mousse birthday cake", etc. according to different ingredients, and generate the third-level code, i.e., the third-level code.

[0092] QR code release

[0093] Customers can independently publish the public and private QR codes generated by this system online and offline.

[0094] Users use QR codes

[0095] After users scan the public QR code, they will be directly directed to the home page of the food producer's information release. They can choose to browse the classified information on the home page according to their preferences. When users scan the "cake" sub-QR code, they will be directly directed to all the information classifications under the "cake" directory of the producer. When users scan the "birthday cake" secondary sub-QR code, they can learn about all the products under the "birthday cake" classification. When users scan the "fruit birthday cake" tertiary QR code, they can learn about relevant information such as the ingredients, production date, production process, specifications, and product features of this birthday cake.

[0096] (2) Implementation Case 2: Application of the tourist scenic area information release system

[0097] In order to strengthen management and improve the tourist experience of visitors, Scenic Area B uses this system to manage and release scenic area information in an all-round, systematic, and digital manner.

[0098] Generation of multi-level QR codes

[0099] Generation of public QR codes: The managers of the tourist scenic area provide basic information such as the enterprise name and account ID and complete registration in this system. The system generates a unique public QR code through a hash function. Scanning the public QR code will directly lead to the home page of the scenic area information release. The home page can publish scenic area introductions, main scenic spots, supporting services, tourist notices, etc., or customize other classified information to be published.

[0100] Generation of private QR codes: Classification identifiers are designed under each classified information, such as "scenic spot introduction", "supporting services", "tourist notices", etc. For each classification, the first-level sub-QR code, that is, the sub-QR code, is generated by adding the classification identifier and the hash encryption sequence. For example, a sub-QR code is generated for the "scenic spot introduction" classification, and on this basis, it can be further subdivided into "Scenic Spot 1", "Scenic Spot 2", "Scenic Spot 3", etc. according to the application scenarios to generate the second-level sub-QR code, that is, the secondary sub-QR code. Under "Scenic Spot 1", it can be divided into "Scenic Spot 1 video", "Scenic Spot 1 voice commentary", etc. to generate the third-level code, that is, the tertiary QR code.

[0101] Publication of QR codes

[0102] Customers independently publish the public and private QR codes generated by this system online and offline.

[0103] Users use QR codes

[0104] After users scan the public QR code, they will be directly directed to the home page of the scenic area information release. They can choose to browse the classified information on the home page according to their preferences. When users scan the "scenic spot introduction" sub-QR code, they will be directly directed to all the information classifications under the "scenic spot introduction" directory of the scenic area. When users scan the "Scenic Spot 1" secondary sub-QR code, they can learn about the detailed information of "Scenic Spot 1". When users scan the "Scenic Spot 1 voice commentary" tertiary QR code, they can further experience the personalized service of Scenic Spot 1.

[0105] Implementation Case 3: Application of Personal Information Publishing System

[0106] This system also provides a package of systematic solutions for individual customers with information publishing needs.

[0107] Multi-level QR code generation

[0108] Public code generation: After an individual's real-name registration, a unique public code is generated. Scanning the public code directly leads to the personal information publishing home page. On the home page, personal profiles, professional skills, hobbies, etc. can be published, or other classified information that customers customize to publish can be provided.

[0109] Private code generation: Individual customers can independently publish relevant information according to their own wishes and publish it according to the system's guidance and classification. For example, "reading", "watching movies", "photography", etc. under "hobbies", to digitalize personal business cards in the form of QR codes.

[0110] QR code publishing

[0111] Customers can independently publish the public and private QR codes generated by this system online and offline.

[0112] Users use the QR code;

[0113] After users scan the public code, they directly reach the personal information publishing home page and can browse the classified information on the home page according to their hobbies; view the corresponding information.

Claims

1. An information publishing system based on big data and hierarchical management of QR codes, characterized by: It includes a multi-level QR code generation engine, a user behavior data collection module, and an intelligent recommendation module; The multi-level QR code generation engine is used to generate a public code and a dynamic hierarchical private code based on the public code according to the customer registration information; User behavior data collection module, which is used to record the user's scanning path in real time, including device type, scanning level and page dwell time, and build user behavior tags; The intelligent recommendation module adopts a dynamic recommendation strategy based on user portraits and user behavior tags, makes single-category recommendations to anonymous users and multi-category recommendations to real-name users, and displays recommended content on PC and mobile terminals.

2. The information publishing system based on big data and hierarchical management of two-dimensional codes according to claim 1 is characterized in that: The generation of the public code according to the customer registration information and the dynamic hierarchical private code based on the public code includes: The public code is a two-dimensional code generated according to user information, and the dynamic hierarchical private code includes multiple layers of subcodes, each layer of subcodes is generated based on the upper layer of subcodes, and the subcode is a two-dimensional code formed by the public code plus a classification identifier plus a hash encryption sequence.

3. The information publishing system based on big data and hierarchical management of two-dimensional codes according to claim 1 is characterized in that: The user behavior tags include interest intensity grading based on page dwell time and category preferences associated with the level of private code scanning.

4. The information publishing system based on big data and hierarchical management of two-dimensional codes according to claim 3 is characterized in that: The interest intensity classification based on page dwell time includes code scanning path analysis, user behavior chain construction and interest weight calculation; The scanning path analysis is performed according to the scanning sequence within the same session to obtain the scanning path; based on the information of each sub-code and the dwell time on the scanning path, a user behavior chain is constructed to obtain the category preference and category interest weight according to the user behavior chain; the category interest weight is obtained by summing up the dwell time of each sub-code under the same public code and calculating the proportion of the dwell time to obtain the interest weight of the category associated with the public code.

5. The information publishing system based on big data and hierarchical management of two-dimensional codes according to claim 4 is characterized in that: The category preference is obtained according to the user behavior chain, wherein the category preference includes single category preference and multi-category preference; the single category preference is that the sub-codes under the same public code are scanned in the same session and the total stay time is greater than the set time; The multi-category preference is that the number of times the sub-codes are scanned across public codes is not less than the set number and the proportion of the stay time under each public code is not less than the set proportion.

6. The information publishing system based on big data and hierarchical management of two-dimensional codes according to claim 5 is characterized in that: The dynamic recommendation strategy based on user portraits and user behavior tags is used to recommend single categories to anonymous users and multiple categories to real-name users, and the recommended content is displayed on PC and mobile terminals, including: The recommendation engine generates a main container, a sub-container and a dynamic sub-container. The main container is bound to a public code to store a customer base portrait and a global recommendation strategy. The sub-container is associated with a sub-code classification to execute a category-level recommendation rule. The dynamic sub-container makes category recommendations based on user behavior tags.

Images