E-commerce anti-user profiling system and method based on artificial intelligence and blockchain

Abstract

The application discloses an e-commerce anti-user portrait system and method based on artificial intelligence and blockchains, comprising a data invasion containment module, a fuzzy user portrait module, an audit recommendation algorithm module and an intelligent comprehensive price comparison module; the data invasion containment module uses blockchain technology to clearly define the data ownership of each node, and decentralizes the data distribution to return the data ownership to the individual user, and desensitizes the authorized user data; the fuzzy user portrait module uses artificial intelligence to automatically search for irrelevant type goods or even opposite type goods; the audit recommendation algorithm module automatically audits the personalized recommendation algorithm, checks the fairness of the algorithm, and manually handles user complaints by accessing an expert system; the intelligent comprehensive price comparison module compares the prices of different e-commerce platforms and the historical prices of the same e-commerce platform, and integrates and displays the price comparison results. The application can effectively solve the problems of personal information leakage and big data killing in the e-commerce field, and improve user satisfaction.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of electronic commerce, and particularly relates to an electronic commerce anti-user portrait system and method based on artificial intelligence and blockchains. BACKGROUND

[0002] Under the background of rapid development of electronic commerce, due to the lack of effective supervision, personal information leakage, big data killing familiar people and other problems gradually stand out. The excessive collection of user data by e-commerce platforms has caused adverse effects. By collecting user information and studying user consumption behavior, a user portrait can be clearly drawn, which is harmful to user information security. This behavior may cause single push and affect the consumer experience, or cause information leakage and threaten the safety of user property and person. In addition, e-commerce platforms have the problem of price discrimination against different user groups. On the one hand, in order to attract new users, new users are provided with preferential prices; on the other hand, using the user portrait and consumption stickiness of old users, the old users are hidden from preferential prices and the prices are raised, and big data killing familiar people is carried out. This unfair sales behavior forces users to bear high prices, disrupts the fair competition order and exacerbates market monopoly.

[0003] Data is the key to personalized recommendation and differentiation of new and old users by e-commerce platforms. Blockchain technology can effectively curb data invasion, and based on distributed ledgers, user data can be decentralized and stored locally. Artificial intelligence technology can use computers and machines to simulate human thinking to solve and make decisions about specific problems. Based on the anti-user portrait of forming a user portrait and conducting differential pricing, artificial intelligence technology can achieve fuzzy user portrait, audit recommendation algorithms and intelligent comprehensive comparison. The above technologies provide a technical basis for protecting user data security and anti-user portrait. SUMMARY

[0004] The purpose of the present application is to provide an electronic commerce anti-user portrait system and method based on artificial intelligence and blockchains, which solves the problems of personal information leakage, single personalized recommendation and big data killing familiar people.

[0005] Technical solution: The present application provides an electronic commerce anti-user portrait system based on artificial intelligence and blockchains, which comprises a data invasion suppression module, a fuzzy user portrait module, an audit recommendation algorithm module, an intelligent comprehensive comparison module, a personalized recommendation system, a real-time database, a historical database, a comparison model library, an expert system and a supervision platform.

[0006] The data invasion suppression module decentralizes and distributes user consumption data and desensitizes it, and stores the desensitized data in the historical database.

[0007] The fuzzy user portrait module processes user consumption data in the historical database, automatically searches for unrelated and opposite categories of goods based on artificial intelligence, and generates a fuzzy user portrait. The fuzzy user portrait information is stored in the historical database.

[0008] The audit recommendation algorithm module automatically audits the algorithm program of the e-commerce platform based on artificial intelligence, and examines the fairness of the algorithm.

[0009] The personalized recommendation system uses the audited recommendation algorithm to calculate personalized recommended goods based on the consumption data in the historical database, and inputs the goods information into the intelligent comprehensive comparison module.

[0010] The intelligent comprehensive comparison module obtains specific information of goods from the real-time database and the historical database according to the calculation results of the e-commerce platform personalized recommendation system, and integrates and displays the comparison results.

[0011] The real-time database stores user-provided information, browsing records and purchase records, and real-time price information of goods displayed on the e-commerce platform.

[0012] The historical database stores desensitized user information, fuzzy user portrait, and historical price information of goods displayed on each e-commerce platform.

[0013] The comparison model library can be called by the intelligent comprehensive comparison module, and then the detailed information of goods on each e-commerce platform is used for comprehensive comparison.

[0014] Further, the system further comprises a visualization panel connected to the intelligent comprehensive comparison module, which visualizes and integrates the comprehensive comparison results.

[0015] Further, the audit recommendation algorithm module is connected to an expert system and a supervision platform, manually handles algorithm unfairness complaints from users, and is subject to supervision by regulatory agencies at any time.

[0016] Further, the intelligent comprehensive comparison module accesses the real-time database to collect price information of the same goods on different e-commerce platforms, and accesses the historical database to collect historical price information of the same goods on the same e-commerce platform.

[0017] Further, the data invasion suppression module uses blockchain technology to clearly define the data ownership of each node, and returns the data ownership to the user and distributes it in the user's local storage.

[0018] Further, the data used by the data invasion suppression module is sourced from the real-time database.

[0019] Based on the same inventive concept, the present application also proposes an e-commerce anti-user profiling method based on artificial intelligence and blockchain. The method uses artificial intelligence and blockchain technology to perform anti-invasion processing on user data, fuzzy processing on user profiles, and auditing of e-commerce platform personalized recommendation algorithms. Based on user autonomy and algorithm fairness, the method provides personalized recommendation services and intelligent comprehensive comparison services for users. Specifically, the method includes the following steps:

[0020] (1) The user inputs personal information and performs browsing and purchasing on the e-commerce platform. The personal information and browsing and purchasing records are recorded in a real-time database.

[0021] (2) The information in the real-time database is decentralized and stored in the user's local device.

[0022] (3) If the user authorizes the use of the data, the relevant information is desensitized and recorded in a historical database, and step (4) is entered. If the user does not authorize, step (9) is entered.

[0023] (4) If the user chooses to perform a fuzzy user profiling operation, the e-commerce platform automatically searches for goods that are not related or even inversely related to the browsing and purchasing records, and records the fuzzy user profile in the historical database, and then enters step (5). If the user does not choose this operation, step (5) is directly entered.

[0024] (5) The fairness of the e-commerce platform's personalized recommendation algorithm is automatically audited.

[0025] (6) If the user complains about the unfairness of the algorithm, the expert system is accessed and the user's complaint is manually handled. If the user does not complain, step (7) is directly entered.

[0026] (7) The supervision platform checks whether there is a problem with the recommendation algorithm module. If there is no problem, step (8) is entered. If there is a problem, step (11) is directly entered.

[0027] (8) The personalized recommendation system uses the audited recommendation algorithm to calculate personalized recommended goods based on the consumption data in the historical database.

[0028] (9) The comparison results are integrated and displayed for the goods recommended to the user or actively browsed by the user.

[0029] (10) The user makes a purchase decision and enters step (1) again.

[0030] (11) The process ends.

[0031] Beneficial effects: compared with the prior art, the beneficial effects of the present application: the user data of the present application is decentralized and desensitized, avoiding illegal occupation of user data by e-commerce platforms; providing a blurred user portrait function, automatically searching for unrelated and opposite goods based on artificial intelligence, avoiding the formation of specific consumer characteristics; the personalized recommendation algorithm accepts internal audit and external supervision, improving the fairness of the algorithm; based on the audited recommendation algorithm and the blurred user portrait, personalized recommendation and comprehensive comparison are carried out, avoiding e-commerce platform differential pricing; solving the problems of personal information leakage, single personalized recommendation and big data killing. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 It is an anti-user portrait system for e-commerce based on artificial intelligence and blockchain.

[0033] Figure 2 It is a flowchart of an anti-user portrait method for e-commerce based on artificial intelligence and blockchain. DETAILED DESCRIPTION

[0034] The present application will be further described in detail below in combination with the drawings.

[0035] The present application provides an anti-user portrait system for e-commerce based on artificial intelligence and blockchain, as shown in Figure 1 The system includes a data invasion suppression module, a blurred user portrait module, an audited recommendation algorithm module, an intelligent comprehensive comparison module, a personalized recommendation system, a real-time database, a historical database, a comparison model library, an expert system and a supervision platform. The user data is decentralized and distributed by using blockchain technology; the blurred user portrait, audited recommendation algorithm and intelligent comprehensive comparison are realized by using artificial intelligence; thereby the user is disguised in the e-commerce level, solving the problems of personal information leakage, single personalized recommendation and big data killing.

[0036] The data invasion suppression module decentralizes and distributes the user consumption data and performs desensitization. The data ownership of each node is determined by using blockchain technology, and the data ownership is returned to the user personally and distributed in the user's local; the e-commerce platform can only obtain the use right of desensitized data (the data is transformed by desensitization rules to realize reliable protection of sensitive private data such as personal information of ID number, mobile phone number and bank card number), and the user's authorization must be obtained before use. The data used in this module is derived from the real-time database, including personal information provided by the user and information such as user browsing and consumption records obtained by the e-commerce platform background. In addition, the desensitized data will be stored in the historical database.

[0037] The fuzzy user portrait module processes the user consumption data in the historical database. With the authorization of the user, the module calls the user consumption data in the historical database, automatically searches irrelevant and opposite goods based on artificial intelligence, and thus fuzzes the user portrait. The fuzzy user portrait information is stored in the historical database, so that the accuracy of the personalized recommendation of the e-commerce platform is reduced.

[0038] The audit recommendation algorithm module audits the algorithm program of the e-commerce platform and examines the fairness of the algorithm. Based on artificial intelligence, the fairness of the recommendation algorithm is automatically examined; through the external expert system, the algorithm unfairness problem complained by the user is manually processed; the audited recommendation algorithm is formed and used for subsequent personalized recommendation of the e-commerce platform. At the same time, the audit recommendation algorithm module is externally connected to the supervision platform and is subject to the examination of the supervision agency at any time.

[0039] The personalized recommendation system uses the audited recommendation algorithm, calculates the personalized recommended goods according to the consumption data in the historical database, and inputs the goods information into the intelligent comprehensive comparison module.

[0040] The intelligent comprehensive comparison module obtains the specific information of the goods from the relevant database according to the calculation result of the e-commerce platform personalized recommendation system, and integrates and displays the comparison result. The real-time database is accessed to collect the price information of the same goods on different e-commerce platforms; the historical database is accessed to collect the historical price information of the same goods on the same e-commerce platform; the comparison model is used for comprehensive comparison, and the integrated display is performed through the visual panel.

[0041] The real-time database stores the information provided by the user, the browsing record and the purchase record, and the real-time price of the goods displayed by the e-commerce platform. The historical database stores the desensitized user information, the fuzzy user portrait, the historical price of the goods displayed by each e-commerce platform, and the like. The comparison model library can be called to perform comprehensive comparison using the detailed information of the goods of each e-commerce platform.

[0042] The expert system provides decision support for manually processing the algorithm unfairness problem complained by the user. The supervision platform is externally connected to the audit recommendation algorithm module, and can supervise the working state of the audit recommendation algorithm module and check the historical audit record at any time. The visual panel is externally connected to the intelligent comprehensive comparison module, and can visualize and integrate the display of the comprehensive comparison result.

[0043] Based on the same inventive concept, the application also provides an e-commerce anti-user portrait method based on artificial intelligence and block chain. The user data is processed by anti-invasion, the user portrait is processed by fuzzing, and the personalized recommendation algorithm of the e-commerce platform is audited, and thus the personalized recommendation system based on user autonomy and algorithm fairness provides goods recommendation service and intelligent comprehensive comparison service for the user, and the effect of anti-user portrait is achieved. Figure 2As shown, specifically comprising the following steps:

[0044] (1) The user inputs personal information and makes purchases on the e-commerce platform, and the personal information and browsing and purchasing records are entered into the real-time database;

[0045] (2) The information in the real-time database is decentralized and stored in the user's local;

[0046] (3) If the user authorizes the data use permission, the relevant information is desensitized and entered into the historical database, and step (4) is entered; if not, it jumps to step (9);

[0047] (4) If the user selects to perform a fuzzy user portrait operation, the e-commerce platform automatically searches for goods that are not related to the browsing and purchasing records or even inversely related, and enters the fuzzy user portrait into the historical database, and enters step (5); If the user does not select this operation, it directly enters step (5);

[0048] (5) Automatically audit the fairness of the e-commerce platform personalized recommendation algorithm;

[0049] (6) If the user complains about the unfairness of the algorithm, access the expert system to manually handle user complaints, and enter step (7); If the user does not complain, it directly enters step (7);

[0050] (7) The supervision platform checks whether the recommendation algorithm module has problems, if not, it enters the next step (8); if there are problems, it directly jumps to step (11);

[0051] (8) The personalized recommendation system uses the audited recommendation algorithm to calculate personalized recommended goods based on the consumption data in the historical database;

[0052] (9) For the goods recommended to the user or actively browsed by the user, integrated display the comparison results;

[0053] (10) The user makes a purchase decision and enters step (1) in a loop;

[0054] (11) End.

[0055] The e-commerce anti-user portrait method based on artificial intelligence and blockchain technology uses artificial intelligence and blockchain technology to perform anti-invasion processing on user data, fuzzy processing on user portraits, and auditing of e-commerce platform personalized recommendation algorithms, and then provides goods recommendation services and intelligent comprehensive comparison services for users based on user autonomy and algorithm fairness. Personalized recommendation system, which can achieve the effect of anti-user portrait.

Claims

1. An anti-user profiling system for e-commerce based on artificial intelligence and blockchain, characterized in that, The system comprises a data invasion containment module, a user profile blurring module, an audit recommendation algorithm module, an intelligent comprehensive comparison module, a personalized recommendation system, a real-time database, a historical database, a comparison model library, an expert system, and a regulatory platform. The data invasion containment module decentralizes and desensitizes user consumption data, and stores the desensitized data in the historical database. The user profile blurring module processes user consumption data in the historical database, automatically searches for unrelated and opposite goods based on artificial intelligence, blurs the user profile, and stores the blurred user profile information in the historical database. The audit recommendation algorithm module automatically audits the algorithm program of the e-commerce platform based on artificial intelligence, and examines the fairness of the algorithm. The personalized recommendation system uses the audited recommendation algorithm to calculate personalized recommended goods based on the consumption data in the historical database, and inputs the goods information into the intelligent comprehensive comparison module. The intelligent comprehensive comparison module obtains specific information of goods from the real-time database and the historical database based on the calculation results of the e-commerce platform personalized recommendation system, and integrates and displays the comparison results. The real-time database stores user-provided information, browsing records, and purchase records, as well as real-time price information of goods displayed on the e-commerce platform. The historical database stores desensitized user information, blurred user profiles, and historical price information of goods displayed on each e-commerce platform. The comparison model library can be called by the intelligent comprehensive comparison module, and then used to conduct comprehensive comparison of detailed information of goods on each e-commerce platform. 2.The electronic commerce anti-user profiling system based on artificial intelligence and blockchain of claim 1, wherein, The system also comprises a visualization panel connected to the intelligent comprehensive comparison module, which visualizes and integrates the comprehensive comparison results. 3.The electronic commerce anti-user profiling system based on artificial intelligence and blockchain of claim 1, wherein, The audit recommendation algorithm module is connected to the expert system and the regulatory platform, manually handles user complaints about algorithm unfairness, and accepts supervision from regulatory agencies at any time. 4.The electronic commerce anti-user profiling system based on artificial intelligence and blockchain of claim 1, wherein, The intelligent comprehensive comparison module accesses the real-time database to collect price information of the same goods on different e-commerce platforms, and accesses the historical database to collect historical price information of the same goods on the same e-commerce platform. 5.The electronic commerce anti-user profiling system based on artificial intelligence and blockchain of claim 1, wherein, The data invasion containment module uses blockchain technology to clearly define the data ownership of each node, and returns the data ownership to the user personally and distributes it in the user's local storage. 6.The electronic commerce anti-user profiling system based on artificial intelligence and blockchain of claim 1, wherein, The data used by the data invasion containment module comes from the real-time database.

7. An AI and blockchain based e-commerce anti-profiling method using the system according to any one of claims 1-6, characterized in that, Using artificial intelligence and blockchain technology, user data is processed against invasion, user profiles are blurred, and e-commerce platform personalized recommendation algorithms are audited. Based on user autonomy and algorithm fairness, personalized recommendation services and intelligent comprehensive comparison services are provided for users. Specifically, the steps include: (1) The user inputs personal information on the e-commerce platform and browses and purchases, and the personal information and browsing and purchase records are recorded in the real-time database; (2) The information in the real-time database is decentralized and stored in the user's local storage; (3) If the user authorizes data usage rights, the relevant information is desensitized and recorded in the historical database, and step (4) is entered; if not, step (9) is entered; (4) If the user chooses to perform a fuzzy user portrait operation, the e-commerce platform automatically searches for goods that are not related or even inversely related to the browsing and purchase records, and enters the fuzzy user portrait into the historical database, and then enters step (5); if the user does not choose this operation, it directly enters step (5); (5) Automatically audit the fairness of the e-commerce platform's personalized recommendation algorithm; (6) If the user complains about the unfairness of the algorithm, access the expert system to manually handle user complaints; if the user does not complain, it directly enters step (7); (7) The supervision platform checks whether there is a problem with the recommendation algorithm module, if not, it enters step (8); if there is a problem, it directly jumps to step (11); (8) The personalized recommendation system uses the audited recommendation algorithm to calculate personalized recommended goods based on the consumption data in the historical database; (9) For the goods recommended to the user or actively browsed by the user, integrate and display the comparison results; (10) The user makes a purchase decision and enters step (1) in a loop; (11) End.

Images